CN118015002A - Traffic engineering road condition visual detection method and system - Google Patents
Traffic engineering road condition visual detection method and system Download PDFInfo
- Publication number
- CN118015002A CN118015002A CN202410425015.6A CN202410425015A CN118015002A CN 118015002 A CN118015002 A CN 118015002A CN 202410425015 A CN202410425015 A CN 202410425015A CN 118015002 A CN118015002 A CN 118015002A
- Authority
- CN
- China
- Prior art keywords
- gray level
- road condition
- defect
- gray
- traffic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 44
- 230000000007 visual effect Effects 0.000 title claims abstract description 28
- 230000007547 defect Effects 0.000 claims abstract description 236
- 238000000034 method Methods 0.000 claims abstract description 66
- 238000003384 imaging method Methods 0.000 claims abstract description 53
- 230000002708 enhancing effect Effects 0.000 claims abstract description 5
- 238000004458 analytical method Methods 0.000 claims description 35
- 230000006870 function Effects 0.000 claims description 14
- 230000011218 segmentation Effects 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 4
- 238000011179 visual inspection Methods 0.000 claims 8
- 230000008859 change Effects 0.000 description 14
- 230000008569 process Effects 0.000 description 10
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000013527 convolutional neural network Methods 0.000 description 8
- 239000010426 asphalt Substances 0.000 description 7
- 238000012549 training Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 230000009467 reduction Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000003064 k means clustering Methods 0.000 description 3
- 230000007704 transition Effects 0.000 description 3
- 230000007423 decrease Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000002310 reflectometry Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000002146 bilateral effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the technical field of image enhancement, in particular to a traffic engineering road condition visual detection method and system. According to the method, a distribution defect characteristic value of the gray level is obtained according to the distribution difference between the gray level and the number of pixel points corresponding to the gray level around the gray level and the proximity degree between the gray level and the gray level of a normal road surface; acquiring an imaging defect rule value of the gray level according to the correlation between all the clustering clusters corresponding to the gray level and the reference lane line connected domain and the distribution defect characteristic value of the gray level; further obtaining a defect probability value of the gray level; acquiring a traffic road condition enhancement image according to defect probability values of all gray levels of an original traffic road condition image; and finally, carrying out defect detection on the traffic road condition enhanced image. The invention improves the accuracy of detecting the visual defects of traffic engineering road conditions by effectively enhancing the prominence of the rut defects.
Description
Technical Field
The invention relates to the technical field of image enhancement, in particular to a traffic engineering road condition visual detection method and system.
Background
Asphalt pavement is a typical flexible pavement. Because asphalt pavement is relatively soft, various defects such as ruts, cracks, grooves and the like are easy to occur on the pavement when the traffic load of the pavement is continuously increased. The visual detection of traffic engineering road conditions is beneficial to road maintenance personnel to accurately know the defect condition of the asphalt pavement, and corresponding maintenance plans are formulated according to the defect condition of the road to ensure driving safety. In recent years, by training a large amount of road image data and utilizing a deep learning model to automatically identify road surface defect condition characteristics, the accurate detection of traffic engineering road condition defects is realized.
In order to enhance the defect expression level, the road surface in the road condition image needs to be enhanced, the image is enhanced by a histogram equalization method in the prior art, and the histogram equalization method is used for redistributing the pixel values of the image, so that the brightness distribution of the image is more uniform, and the contrast and detail of the defect in the image are enhanced. However, under the condition of no lower-angle light irradiation, the rut characteristic expression degree is relatively weak, and the histogram equalization method ignores the rut defect probability of the gray level, so that the contrast ratio of the gray level with high defect probability and the gray level with low defect probability cannot be effectively amplified, the prominence degree of the rut defect cannot be effectively enhanced, and the traffic engineering road condition visual defect detection is inaccurate.
Disclosure of Invention
In order to solve the technical problems that the prior art uses a histogram equalization method to enhance images and cannot effectively enhance the prominence degree of rutting defects, so that the traffic engineering road condition visual defect detection is inaccurate, the invention aims to provide a traffic engineering road condition visual detection method and a system, and the adopted technical scheme is as follows:
a traffic engineering road condition visual detection method, comprising the following steps:
Acquiring an original image of traffic road conditions;
Acquiring a road condition gray level histogram of the traffic road condition original image; acquiring normal road surface gray levels according to the number of pixel points corresponding to all gray levels in the road condition gray level histogram; in the road condition gray level histogram, according to the distribution difference between the gray level and the number of pixels corresponding to the surrounding gray level and the proximity degree between the gray level and the normal road surface gray level, obtaining a distribution defect characteristic value of the gray level;
Performing binarization processing on the traffic road condition original image to obtain a reference lane line connected domain; clustering all pixel points corresponding to the gray level in the traffic road condition original image to obtain a plurality of clustering clusters corresponding to the gray level; acquiring an imaging defect rule value of the gray level according to the correlation between all the clustering clusters corresponding to the gray level and the reference lane line connected domain and the distribution defect characteristic value of the gray level;
Acquiring a defect probability value of the gray level according to the distribution of the road condition gray level histogram, an imaging defect rule value of the gray level and a distribution defect characteristic value of the gray level; according to the defect probability values of all gray levels of the traffic road condition original image, enhancing the traffic road condition original image to obtain a traffic road condition enhanced image;
And performing defect detection on the traffic road condition enhanced image.
Further, the method for acquiring the characteristic value of the distributed defect specifically includes:
obtaining a variation characteristic parameter value according to a variation characteristic parameter value formula, wherein the variation characteristic parameter value formula comprises:
;/> in the road condition gray level histogram, the first/> Variable characteristic parameter values for the individual gray levels; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)Is an absolute value symbol; /(I)A first adjustment factor that is denominator;
performing curve fitting according to the road condition gray level histogram to obtain a fitted gray level curve; segmenting a gray level sequence formed by gray levels of the road condition gray level histogram to obtain each gray level segment;
Obtaining a distribution defect characteristic value according to a distribution defect characteristic value formula, wherein the distribution defect characteristic value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)Is the gray level of a normal road surface; /(I)In the road condition gray level histogram, the first/>Variable characteristic parameter values for the individual gray levels; /(I)To at/>In the gray scale section to which each gray level belongs, all gray levels correspond to the variance of the derivative on the fitted gray scale curve; /(I)A second regulator that is denominator; /(I)In natural number/>Is an exponential function of the base.
Further, the method for acquiring the regular value of the imaging defect specifically includes:
Acquiring a reference lane contour of a reference lane line connected domain; splitting the reference lane contour to obtain all the reference lane sectional contours of the reference lane contour; taking the longest reference lane segmentation contour as a reference driving contour edge; acquiring a reference driving characteristic sequence of the reference driving contour edge according to the distribution of pixel points in the communication domain of the reference driving contour edge and the reference lane line;
Obtaining a cluster contour and a cluster area of each cluster corresponding to each gray level by using a convex hull detection method; splitting each class cluster contour to obtain all class cluster segmentation contours of each class cluster contour; acquiring a running characteristic sequence to be analyzed of each cluster segmentation contour according to each cluster segmentation contour and the corresponding pixel point distribution in the cluster region;
And acquiring imaging defect rule values of gray levels according to the correlation between the reference driving characteristic sequence and the driving characteristic sequence to be analyzed of each cluster sectional profile, the distribution defect characteristic values of gray levels and the fluctuation of the gray levels corresponding to the lengths of all cluster sectional profiles.
Further, the method for obtaining the regular value of the imaging defect comprises the following steps:
acquiring an imaging defect rule value according to an imaging defect rule value formula, wherein the imaging defect rule value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>Imaging defect rule values for the individual gray levels; /(I)For/>The number of gray levels corresponds to the total number of all class cluster segmented contours; /(I)For/>The gray level corresponds to the/>The length of the individual cluster segmentation profile; /(I)For/>The gray levels correspond to the average value of the lengths of the segmented contours of all the class clusters; /(I)In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)For/>The gray level corresponds to the/>A running characteristic sequence to be analyzed of the sectional profile of each cluster; /(I)A reference travel feature sequence that is a reference travel profile edge; /(I)For/>The individual gray levels correspond to/> of the segmented contours of all class clustersMaximum value of (2); /(I)Is an absolute value symbol; /(I)To take the pearson correlation coefficient function.
Further, the method for acquiring the reference driving characteristic sequence and the driving characteristic sequence to be analyzed comprises the following steps:
Taking any one of the reference driving contour edges or the cluster-like segmented contours as a target analysis edge; taking any cluster area or a reference lane line connected area as a target analysis area; taking all the pixel points in the target analysis area as the pixel points of the target area; taking all pixel points on the target analysis edge as target edge pixel points;
in the target analysis area to which the target analysis edge belongs, sequentially counting the total number of all target area pixel points of the target edge pixel points in a preset direction by taking the end point of any one of the target analysis edges as a starting point and the extending direction of the target analysis edge as a direction, so as to obtain a target running characteristic sequence of the target analysis edge;
And taking the target running characteristic sequence of the reference running contour edge as the reference running characteristic sequence of the reference running contour edge, and taking the target running characteristic sequence of the cluster-like segmented contour as the running characteristic sequence to be analyzed of the cluster-like segmented contour.
Further, the method for obtaining the defect probability value comprises the following steps:
obtaining a defect probability value according to a defect probability value formula, wherein the defect probability value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>-Said defect probability values for individual gray levels; /(I)In the road condition gray level histogram, the first/>Imaging defect rule values for the individual gray levels; /(I)In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)The deviation of the road condition gray level histogram is obtained; /(I)The kurtosis of the road condition gray level histogram; /(I)In natural number/>Is an exponential function of the base.
Further, the method for acquiring the normal pavement gray level comprises the following steps:
And taking the gray level corresponding to the maximum number of pixels in the road condition gray level histogram as the normal road surface gray level.
Further, the method for acquiring the reference lane line connected domain specifically includes:
Based on an Ojin threshold method, carrying out binarization processing on the traffic road condition original image to obtain a traffic binarization image; and extracting each front Jing Liantong domain of the traffic binarization image by using a connected domain analysis algorithm, and taking the largest front Jing Liantong domain as a reference lane line connected domain.
Further, the method for acquiring the traffic road condition enhanced image comprises the following steps:
Acquiring an initial mapped gray level of the gray level based on a histogram equalization method;
Acquiring updated mapped gray levels of each gray level according to the defect probability value corresponding to the gray level and the initial mapped gray level of the gray level; the defect probability value and the updated mapped gray level show positive correlation; the initial mapped gray level and the updated mapped gray level exhibit a positive correlation;
And based on a histogram equalization method, acquiring a traffic road condition enhanced image according to the updated mapped gray level of each gray level.
The invention provides a traffic engineering road condition visual detection system which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the steps of any traffic engineering road condition visual detection method when executing the computer program.
The invention has the following beneficial effects:
Considering that the rut defect is a slight depression with regular shape and is represented as a rut texture area formed by a normal reflection area of a pavement and a shadow area caused by deformation in an image, however, under the condition of poor light conditions, the rut texture is not obvious, and the image enhancement is required to be carried out on the rut defect.
Firstly, taking the fact that the gray value consistency of a normal road surface is high and the gray value of the normal road surface is high into consideration, and acquiring the gray level of the normal road surface; the normal road surface gray level reflects the gray level of the normal road surface; considering that the rut texture is due to the light and shadow change generated by rut depression, the reflectivity of the asphalt highway pavement is weak, and the gray level of the rut texture is close to the gray level of the normal pavement; meanwhile, the fact that the rut texture is an area formed by a normal reflection area of the pavement and a shadow area caused by deformation in the image due to pavement deformation is considered, the gray level of the asphalt pavement is uniform, the transition is gentle, and the distribution defect characteristic value of the gray level is obtained; the distribution defect characteristic value is used for preliminarily determining the probability of the rut defect by analyzing the number of the pixel points corresponding to the gray level.
Because the gray level of the lane line is far greater than the gray level of the pavement part in the road condition image, the original image of the traffic road condition is subjected to binarization processing to obtain a reference lane line connected domain; the reference lane line communication domain may reflect a driving direction of the driver. In order to analyze the graph distribution characteristics corresponding to the gray level, clustering is carried out on all pixel points corresponding to the gray level, and a plurality of clustering clusters corresponding to the gray level are obtained; the cluster can reflect the pixel points with similar distances in the same gray level, so as to reflect the distribution of the pixel points in the gray level in the actual position. Considering that the shape of the rut is generally long, the rut and the lane line can reflect the running direction of the vehicle, so that the outline of the rut and the lane line have similarity; acquiring an imaging defect rule value of a determined gray level through the relevance of the actual distribution rule of the pixel points of the gray level in the image and the change rule of the pixel points of the lane line; the larger the imaging defect rule value of the gray level is, the larger the probability that the gray level in the image is the rut gray level is.
The distribution defect characteristic value determines the probability that the gray level is the defect gray level according to the distribution of the number of pixel points in the road condition gray level histogram; the imaging defect rule value mainly determines the probability that the gray level is the defect gray level through the actual distribution rule in the image; obtaining a defect probability value of the gray level by further combining the distribution of the gray level histogram; the defect probability value may more accurately reflect the probability that the gray level is a rut defect gray level. According to the probability that the gray level is the rut texture gray level, the image is adaptively enhanced, and the traffic road condition enhancement image capable of reflecting rut defects more accurately is determined, so that the defect type detection result of the traffic road condition enhancement image is more accurate.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a traffic engineering road condition visual detection method according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of a traffic engineering road condition visual detection method and system according to the invention, which are specific embodiments, structures, features and effects thereof, with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a traffic engineering road condition visual detection method and a system thereof, which are specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flow chart of a traffic engineering road condition visual detection method according to an embodiment of the invention is shown, the method includes the following steps:
step S1, an original image of the traffic road condition is obtained.
The invention is mainly aimed at enhancing the prominence degree of rut defects, so that an original image of traffic road conditions is needed to be acquired first. Specifically, road condition images are acquired through a road detection vehicle, and road condition initial images are acquired; because the road detection vehicle can be generally provided with other detection equipment to evaluate the road condition, the image features are unclear due to the electronic noise interference in the processes of image acquisition and data transmission, so that the noise reduction operation is carried out on the road condition initial image, and the noise reduction image is obtained. The noise reduction operation can eliminate the influence caused by noise and external interference, and the accuracy of subsequent analysis is enhanced. The image completely contains the current road surface information, wherein the color information contains fewer characteristics, the noise reduction image is subjected to gray processing, and the original image of the traffic road condition is obtained, so that rut defects can be conveniently analyzed. The embodiment of the invention adopts bilateral filtering to reduce noise of the image, and an implementer can set the image according to actual conditions.
It should be noted that, in order to facilitate the operation, all index data involved in the operation in the embodiment of the present invention is subjected to data preprocessing, so as to cancel the dimension effect. The specific means for removing the dimension influence is a technical means well known to those skilled in the art, and is not limited herein.
Step S2, obtaining a road condition gray level histogram of an original image of the traffic road condition; acquiring normal road surface gray levels according to the number of pixel points corresponding to all gray levels in the road condition gray level histogram; and in the road condition gray level histogram, acquiring a gray level distribution defect characteristic value according to the distribution difference between the gray level and the number of pixels corresponding to the surrounding gray level and the proximity degree between the gray level and the normal road surface gray level.
Considering that the rut defect is a slight depression with regular shape and is represented as a rut texture area formed by a normal reflection area of a pavement and a shadow area caused by deformation in an image, however, under the condition of poor light conditions, the rut texture is not obvious, and the image enhancement is required to be carried out on the rut defect.
Firstly, taking the fact that the gray value consistency of a normal road surface is high and the gray value of the normal road surface is high into consideration, and acquiring the gray level of the normal road surface; the normal road surface gray level reflects the gray level of the normal road surface; considering that the rut texture is due to the light and shadow change generated by rut depression, the reflectivity of the asphalt highway pavement is weak, and the gray level of the rut texture is close to the gray level of the normal pavement; meanwhile, the fact that the rut texture is an area formed by a normal reflection area of the pavement and a shadow area caused by deformation in the image due to pavement deformation is considered, the gray level of the asphalt pavement is uniform, the transition is gentle, and the distribution defect characteristic value of the gray level is obtained; the distribution defect characteristic value is used for preliminarily determining the probability of the rut defect by analyzing the number of the pixel points corresponding to the gray level.
Preferably, in consideration of that the gray level of the normal road surface is uniform and the normal road surface occupies a relatively high proportion in the image, in one embodiment of the present invention, the method for acquiring the gray level of the normal road surface includes:
In the road condition gray level histogram, the maximum number of pixels corresponds to gray level, and the gray level of the normal road surface is used as the gray level of the normal road surface, and the gray level of the normal road surface is reflected.
Specifically, in one embodiment of the invention, a road condition gray level histogram of the traffic road condition original image is constructed according to gray values of all pixels in the traffic road condition original image, the abscissa of the road condition gray level histogram is gray level, and the ordinate of the road condition gray level histogram is the number of pixels. And analyzing the rut defect probability of the pixel points by analyzing the distribution of the number of the pixel points in the road condition gray level histogram.
Preferably, considering that the road surface gray level is relatively uniform, the number of pixels in the histogram should be approximately gaussian, and since the rut defect is represented in the image as a rut texture region consisting of a normal reflection region of the road surface and a shadow region caused by deformation, for the road surface gray level histogram with a certain gaussian property, the number of pixels in the gray level range of the reduced number of pixels is equivalent to that of pixels, that is, the rut texture portion has a certain regular feature on the distribution of the gray level histogram. The step obtains the characteristic value of the distribution defect through the distribution characteristics of different gray levels on the gray level histogram.
In one embodiment of the invention, firstly, the relatively uniform gray level of the pavement is considered, and the rut defect is represented in the image as a rut texture area formed by a normal reflection area of the pavement and a shadow area caused by deformation, so that the gray level is relatively nonuniform. In the road condition gray level histogram, according to the difference between the gray level and the number of corresponding pixels between adjacent gray levels, obtaining a change characteristic parameter value of the gray level, wherein a change characteristic parameter value formula comprises:
;/> in the road condition gray level histogram, the first/> Variable characteristic parameter values for the individual gray levels; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)Is an absolute value symbol; /(I)Is the denominator first adjustment factor. In one embodiment of the present invention, the first denominator adjustment factor is 0.01, and the practitioner can set itself according to the implementation scenario. It should be noted that, considering that the first gray level does not have the previous gray level and the last gray level does not have the next gray level in the road condition gray level histogram, in one embodiment of the present invention, the number of pixels of the previous adjacent gray level of the first gray level and the number of pixels of the next adjacent gray level of the last gray level in the road condition gray level histogram are fitted by using a difference fitting method.
In the formula of the value of the variation characteristic parameter,The pixel point quantity difference between the current gray level and the previous adjacent gray level is reflected; /(I)The pixel point quantity difference between the current gray level and the next adjacent gray level is reflected; The difference of the pixel point quantity difference between the current gray level and the adjacent gray level is reflected; the effect is to/> And (3) performing standardization so that the change characteristic parameter value reflects the relative difference degree of the pixel point quantity difference between the current gray level and the adjacent gray level, and considering that the gray level of the rut defect area is relatively uneven, the larger the relative difference degree is, the larger the change characteristic parameter value is, and the more likely the rut defect gray level is.
In order to analyze the gray level change characteristics in the road condition gray level histogram, firstly, a fitting gray level curve of the road condition gray level histogram needs to be obtained, and the specific process comprises the following steps:
Specifically, a two-dimensional coordinate system based on a gray histogram, namely, the horizontal axis of the two-dimensional coordinate is gray level, the vertical axis is the number of pixel points of the gray level, each gray level and the number of the corresponding pixel points form a data point, and the data points of all gray levels are counted to be used as fitting data points; and performing curve fitting on all fitting data points by using a least square method to obtain a fitting gray scale curve.
In order to analyze the distribution characteristics around the gray level, the horizontal axis of the histogram of the road Kuang Huidu is segmented to obtain each gray level segment, and the specific process includes:
in one embodiment of the invention, derivative corresponding to each gray level on the fitted gray level curve is obtained by deriving the corresponding data point of each gray level; and taking the gray level with the derivative of 0 as each divided gray level, and segmenting a gray level sequence formed by the gray levels of the road Kuang Huidu histogram by using all the divided gray levels to obtain each gray level segment.
In other embodiments of the present invention, the interval of the road condition gray level histogram corresponding to the horizontal axis is equally divided into a preset number of gray level segments; in one embodiment of the present invention, the number of preset dividing segments is 25, and an implementer can set the preset dividing segments according to implementation scenarios.
In one embodiment of the invention, in order to preliminarily determine the probability of rut defects, a distribution defect characteristic value is obtained according to the distribution of the number of pixels in a road condition gray level histogram. The distributed defect characteristic value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)Is the gray level of a normal road surface; /(I)In the road condition gray level histogram, the first/>Variable characteristic parameter values for the individual gray levels; To at/> In the gray scale section to which each gray level belongs, all gray levels correspond to the variance of the derivative on the fitted gray scale curve; A second regulator that is denominator; /(I) In natural number/>Is an exponential function of the base. In one embodiment of the present invention, the denominator second adjustment factor is 0.001, and the practitioner can set itself according to the implementation scenario. It is noted that/>The specific acquisition process of the method comprises the steps of deriving corresponding data points of each gray level on a fitting gray level curve to acquire derivatives corresponding to the gray levels; in/>In the gray level segment of each gray level, calculating the variance of derivatives corresponding to all gray levels to obtain/>。
In the formula of the characteristic value of the distributed defect,Reflect the/>The difference degree of the individual gray level and the normal road surface gray level is larger as the difference degree is smaller and the distribution defect characteristic value is larger, the probability that the gray level is the defect gray level is larger as the gray level of the rut texture is close to the gray value of the normal road surface; /(I)The uniform degree of the overall change of the number of the pixel points in the interval of the gray level section to which the gray level belongs is reflected, the more uniform the overall change difference of the number of the pixel points in the representative interval is, namely the more likely the gray level of the interval represents the gray level of the area with uniform gray level in the image and the more gentle transition area, and the more accords with the distribution rule of the gray values of the highway; /(I)Reflecting the relative difference degree of the pixel point quantity difference between the gray level and the adjacent gray level, and considering that the gray level of the rut defect area is relatively uneven, the larger the relative difference degree is, the larger the distribution defect characteristic value is, and the larger the probability that the gray level is the defect gray level is; will/>As/>And the adjustment factor of the number difference of the adjacent gray level pixels is adjusted, so that the probability that the gray level is the rutting defect gray level is more accurately expressed. The probability of rut defects is primarily determined by analyzing the distribution of the number of pixel points in the road condition gray level histogram, and the larger the distribution defect characteristic value is, the larger the probability that the gray level is the defect gray level is.
Step S3, binarizing the original image of the traffic road condition to obtain a reference lane line connected domain; clustering all pixel points corresponding to the gray level in an original image of the traffic road condition to obtain a plurality of clusters corresponding to the gray level; and acquiring an imaging defect rule value of the gray level according to the correlation between all the clustering clusters corresponding to the gray level and the reference lane line connected domain and the distribution defect characteristic value of the gray level.
In order to avoid the defect probability of gray level analysis by only the road condition gray level histogram is not accurate enough due to the interference of other objects in the image on different gray levels, the probability of the gray level pixel points containing rut missing pixel points is judged by combining the actual position distribution of the gray level pixel points in the original image of the traffic road condition. In an original image of traffic road conditions, according to the imaging principle of near-large and far-small, the imaging area of the ruts close to the sampling point is larger than that of the ruts far away from the sampling point. In the image, there should therefore be a gradual change in the number of pixels of the rut region in the non-rut line direction. To measure this degree of variation, a fixed reference in the image needs to be found for comparison. In the original image of traffic conditions, lane lines refer to lines for guiding a driver to run on a road, track defects are mainly slight depressions generated on the road surface due to repeated rolling of vehicle tires on the road surface, tracks can reflect the running direction of vehicles, lane lines refer to lines for guiding the driver to run on the road and can reflect the running direction of vehicles, so that the outlines of the tracks and the lane lines are similar, and the pixel count amounts of the track areas and the lane line areas have similar change rules.
Preferably, the subsequent analysis of the rut imaging features requires a fixed reference, and the reference lane line connected domain is acquired in consideration of the similar change rule of the number of pixels in the rut region and the lane line region. In one embodiment of the invention, the method for acquiring the reference lane line connected domain specifically comprises the following steps:
Because the gray level of the lane line is far greater than the gray level of the pavement part in the road condition image, the original image of the traffic road condition is subjected to binarization processing based on an Ojin threshold method, and a traffic binarization image is obtained; the foreground pixel points in the traffic binarized image represent the lane line pixel points. And extracting each front Jing Liantong domain of the traffic binarization image by using a connected domain analysis algorithm, and taking the largest front Jing Liantong domain as a reference lane line connected domain. The reference lane line connected domain represents the area corresponding to the longest lane line, and can well reflect the running track of the vehicle, so that the reference lane line connected domain can be used as a fixed reference for analyzing the imaging characteristics of the ruts.
In order to analyze the graph distribution characteristics corresponding to the gray level, clustering is carried out on all pixel points corresponding to the gray level, and a plurality of clustering clusters corresponding to the gray level are obtained; the cluster can reflect the pixel points with similar distances in the same gray level, so as to reflect the distribution of the pixel points in the gray level in the actual position. Considering that the shape of the rut is generally long, the rut and the lane line can reflect the running direction of the vehicle, so that the outline of the rut and the lane line have similarity; acquiring an imaging defect rule value of a determined gray level through the relevance of the actual distribution rule of the pixel points of the gray level in the image and the change rule of the pixel points of the lane line; the larger the imaging defect rule value of the gray level is, the larger the probability that the gray level in the image is the rut gray level is.
Specifically, in order to determine the probability that a rut defect pixel exists in a pixel corresponding to a gray level, it is necessary to analyze that the distribution characteristic of the gray level in an image accords with the rut imaging characteristic, and firstly, the actual position distribution of the pixel of the gray level in the image is determined. In the original image of the traffic road condition, K-means clustering (K-Means Clustering, K mean clustering algorithm) is carried out on all pixel points corresponding to the gray level according to Euclidean distance among the pixel points, and a plurality of clusters corresponding to the gray level are obtained. It should be noted that K-means clustering is a prior art well known to those skilled in the art, and only a brief process of acquiring a plurality of clusters corresponding to gray levels is described herein: in an original image of traffic road conditions, K pixel points are selected as an initial clustering center by using an elbow method, and a distance measurement value of two pixel points is calculated according to Euclidean distance between the two pixel points; clustering is carried out according to the distance metric value from the pixel point to each clustering center point, and a plurality of clustering clusters corresponding to gray levels are obtained. The cluster is the pixel points with the same gray level and relatively close distribution distance, so that the distribution of the pixel points with the gray level in the actual position is reflected.
Preferably, the probability that the gray level is the rut defect gray level is further analyzed through the actual image distribution characteristics of the pixel points corresponding to the gray level, and the imaging defect rule value is obtained. In one embodiment of the present invention, the method for obtaining the regular value of the imaging defect specifically includes:
First, in order to determine an edge reflecting a traveling direction of a vehicle in a reference lane line connected domain, a reference lane contour of the reference lane line connected domain is acquired by a convex hull detection method. In order to determine the edge reflecting the running direction of the vehicle on the contour, the convex hull detection method is utilized to obtain the vertex on the reference lane contour, the reference lane contour is split by utilizing the vertex, and all the reference lane segment contours of the reference lane contour are obtained; considering that the edges in the driving direction of the lane tend to be long and have strong referential property in the reference lane line communication domain, taking the longest reference lane segment profile as the reference driving profile edge; the reference driving profile edge may represent an edge of the driving direction of the vehicle.
In order to analyze the correlation between the formation area represented by the gray-level pixel points and the lane line area, the cluster clusters are pixel points with the same gray level and relatively close distribution distance, so that the distribution of the gray-level pixel points in the actual position can be reflected. Obtaining cluster contour of each cluster corresponding to each gray level by using a convex hull detection method, wherein a region surrounded by the cluster contour is used as a cluster region; the cluster-like profile can reflect the geometric characteristics of the gray level pixel points in the actual distribution, and the gray level pixel points in the cluster-like region are in the actual distribution. And the convex hull detection method can also be used for obtaining the vertexes on the cluster-like contours, and the cluster-like contours are split by using the vertexes to obtain all the cluster-like segmented contours of the cluster-like contours. The cluster-like segmented contour can reflect the geometric local characteristics of the pixel points corresponding to the gray level in actual distribution.
Preferably, in order to analyze the correlation between the actual distribution rule of the gray level pixels in the image and the actual distribution rule of the lane lines in the image, the imaging characteristics of the connected domain of the reference lane lines need to be analyzed to determine the reference driving characteristic sequence, and the imaging characteristics of the gray level in the actual distribution need to be analyzed to determine the driving characteristic sequence to be analyzed. In one embodiment of the present invention, the method for acquiring the reference driving feature sequence and the driving feature sequence to be analyzed includes:
taking any one of the reference driving contour edges or cluster-like segmented contours as a target analysis edge; taking any cluster area or a reference lane line connected area as a target analysis area; taking all pixel points in the target analysis area as the pixel points of the target area; taking all pixel points on the target analysis edge as target edge pixel points;
In a target analysis area to which a target analysis edge belongs, sequentially counting the total number of all target area pixel points of target edge pixel points in a preset direction by taking an endpoint of any one target analysis edge as a starting point and taking the extending direction of the target analysis edge as a direction, so as to obtain a target running characteristic sequence of the target analysis edge;
Taking the target running characteristic sequence of the reference running contour edge as the reference running characteristic sequence of the reference running contour edge; the elements in the reference driving feature sequence may represent the number of pixels in the reference lane line connected domain of the target edge pixels in the preset direction. The reference driving feature sequence can reflect the distribution of pixel points in the lane line in the extending direction of the edge of the reference driving contour, so as to further represent the actual distribution rule of the lane line in the image.
And taking the target running characteristic sequence of the cluster-like segmented contour as a running characteristic sequence to be analyzed of the cluster-like segmented contour. The elements in the running characteristic sequence to be analyzed can represent the number of pixel points of the target edge pixel points in the upper cluster contour. The running characteristic sequence to be analyzed can reflect the distribution of the pixel points in the cluster-like contour in the extending direction of the cluster-like sectional contour, and further represents the actual distribution rule of the gray-level pixel points in the image.
In one embodiment of the present invention, the tangential direction of the target analysis edge at each target edge pixel point is obtained, and the direction perpendicular to the tangential direction is taken as the preset direction of the target edge pixel point, so that the implementer can set the direction according to the implementation scene. In other embodiments of the present invention, a straight line fitting may be performed on the target analysis edge, that is, the reference driving contour edge or the cluster-like segmented contour, to obtain a fitted driving straight line, and a direction perpendicular to the fitted driving straight line is taken as a preset direction, so that an implementer may set the direction according to the implementation scenario.
Considering that the shape of the rut is generally long, the rut and the lane line can reflect the running direction of the vehicle, so that the rut and the lane line have similar outlines and obtain the regular value of the imaging defect of the determined gray level. In one embodiment of the present invention, the imaging defect rule value formula includes:
; wherein/> In the road condition gray level histogram, the first/>Imaging defect rule values for the individual gray levels; /(I)For/>The number of gray levels corresponds to the total number of all class cluster segmented contours; /(I)For/>The gray level corresponds to the/>The length of the individual cluster segmentation profile; /(I)For/>The gray levels correspond to the average value of the lengths of the segmented contours of all the class clusters; /(I)In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)For/>The gray level corresponds to the/>A running characteristic sequence to be analyzed of the sectional profile of each cluster; /(I)A reference travel feature sequence that is a reference travel profile edge; for/> The individual gray levels correspond to/> of the segmented contours of all class clustersMaximum value of (2); /(I)Is an absolute value symbol; /(I)To take the pearson correlation coefficient function.
In the imaging defect rule value formula, considering that the track shape is generally a strip shape, the gray level is larger in the difference of the sectional profiles of all the class clusters,The difference of the lengths of the gray levels corresponding to the sectional outlines of all the class clusters is reflected, the larger the difference is, the larger the probability that the gray level is the rut gray level is, and the larger the imaging defect rule value is; considering that the same gray level corresponds to a plurality of cluster-like segmented contours, the maximum value is selected to select the value with high correlation between the running characteristic sequence to be analyzed and the reference running characteristic sequence of the cluster-like segmented contour as much as possible, so as to analyze the probability of rutting defect pixels existing in pixels corresponding to the gray level,/>The larger the grey level is, the larger the probability that the rut defect pixel exists in the corresponding pixel point is, the larger the probability that the grey level is the rut grey level is, and the larger the imaging defect rule value is; in order to avoid possible interference of other objects on the road surface with the similarity calculation, the method comprises the following steps ofFor this/>Adjusting; the imaging defect rule value reflects the probability that the pixel points with the rut defects exist in the pixel points corresponding to the gray levels, and the larger the imaging defect rule value is, the more likely the gray levels are the rut defect gray levels.
Step S4, according to the distribution of the road condition gray level histogram, the imaging defect rule value of the gray level and the distribution defect characteristic value of the gray level, obtaining the defect probability value of the gray level; and carrying out enhancement processing on the original traffic road condition image according to the defect probability values of all gray levels of the original traffic road condition image to obtain an enhanced traffic road condition image.
Through the steps, the probability that the gray level is the defect gray level is determined by the distribution of the number of the pixel points in the road condition gray level histogram according to the distribution defect characteristic values; the imaging defect rule value mainly determines the probability that the gray level is the defect gray level through the actual distribution rule in the image; obtaining a defect probability value of the gray level by further combining the distribution of the gray level histogram; the defect probability value may more accurately reflect the probability that the gray level is a rut defect gray level. And obtaining a traffic road condition enhancement image according to defect probability values of all gray levels of an original traffic road condition image, and realizing self-adaption aiming at enhancement of the image according to the probability that the gray levels are rut texture gray levels.
Preferably, in order to determine a probability that the gray level can be more accurately reflected as the rut defect gray level, in one embodiment of the present invention, the method for obtaining the defect probability value includes:
Obtaining a defect probability value according to a defect probability value formula, wherein the defect probability value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>Defect probability values for the individual gray levels; /(I)In the road condition gray level histogram, the first/>Imaging defect rule values for the individual gray levels; in the road condition gray level histogram, the first/> Distribution defect feature values of the individual gray levels; /(I)The deviation of the road condition gray level histogram; /(I)The kurtosis of the road condition gray level histogram; /(I)In natural number/>Is an exponential function of the base.
In the formula of the defect probability value,The Gaussian characteristic of the gray level histogram of the road condition is reflected, the smaller the skewness and kurtosis are, the stronger the Gaussian characteristic of the gray level histogram is, the purer the road surface in the representative image is, the less interference is caused by other objects with approximate gray values, the larger the Gaussian characteristic is, and the gravity center can be judged to be more biased to the distribution defect characteristic value; otherwise, determining the center of gravity may be more biased toward the imaging defect law value. By/>The weights of the distribution defect characteristic values and the imaging defect rule values are adjusted, so that the defect probability values can more accurately reflect the probability that the gray level is the rut defect gray level.
Preferably, in the histogram equalization process, the gray values of the image are redistributed, so that the gray value distribution of the image is more uniform, and the contrast and visual effect of the image are enhanced. In order to enhance the highlighting degree of the defective gray level, the gray level mapping is adjusted through defect probability value adjustment, and when the updated mapped gray level of each gray level is obtained, the probability of being distributed to a new gray level after the gray level with larger defect probability is mapped is larger, so that the highlighting degree of the rut defect is effectively enhanced. In one embodiment of the invention, the method for acquiring the traffic road condition enhanced image comprises the following steps:
Acquiring an initial mapped gray level of the gray level based on a histogram equalization method; it should be noted that, obtaining the initial mapped gray level of the gray level by the histogram equalization method is a well known technology of those skilled in the art, where the formula for determining the initial mapped gray level of the gray level by the histogram equalization method is:
; wherein/> For/>An initial mapped gray level of the individual gray levels; Is a rounding function; /(I) Is the total number of gray levels; /(I)First/>The number of pixels for each gray level, a=0, 1,2, … L-1; /(I)The total number of all the pixel points in the original image of the traffic road condition is obtained.
Acquiring updated mapped gray levels of each gray level according to the defect probability value corresponding to the gray level and the initial mapped gray level of the gray level; the defect probability value and the gray level after the updating mapping show positive correlation; the initial mapped gray level and the updated mapped gray level exhibit a positive correlation.
The positive correlation relationship indicates that the dependent variable increases with the increase of the independent variable, and the dependent variable decreases with the decrease of the independent variable, and the specific relationship may be a multiplication relationship, an addition relationship, an idempotent of an exponential function, and is determined by practical application.
In order to enhance the prominence of the defect gray level, the updated mapped gray level is obtained by combining the defect probability value of the gray level. In one embodiment of the present invention, the updated mapped gray level acquisition formula includes:
; wherein/> For/>Updated mapped gray levels of the individual gray levels; /(I)Is a rounding function; /(I)Is the total number of gray levels; /(I)For/>The number of pixels for each gray level; /(I)The total number of all pixel points in the original image of the traffic road condition is the total number of all pixel points in the original image of the traffic road condition; /(I)For/>Defect probability values for the individual gray levels; /(I)The accumulated value of defect probability values of all gray levels in the original image of the traffic road condition is obtained; /(I)For/>An initial mapped gray level of the individual gray levels.
In the updated mapped gray level formula, the updated mapped gray level not only considers the number of pixels corresponding to the gray level, but also considers the defect probability value of the gray level being the rut defect gray level, and the larger the defect probability value is, the more likely the defect probability value is the gray level containing rut defect pixels, so that the larger the updated mapped gray level is, the difference between the defect area and the mapped gray level after the normal area is increased, the contrast of the rut defect part is improved, and the false detection rate or the omission rate is reduced.
And acquiring a traffic road condition enhanced image according to the updated mapped gray level of each gray level. It should be noted that, the histogram equalization method is a well-known prior art for those skilled in the art, and only a brief process of obtaining the traffic road condition enhanced image according to the updated mapped gray level of each gray level is briefly described herein:
And according to the updated mapped gray level of each gray level and in combination with the CLAHE limiting contrast, applying the updated mapped gray level of each gray level to the original traffic road condition image to obtain the enhanced traffic road condition image.
And S5, performing defect detection on the traffic road condition enhanced image.
By the steps, the traffic road condition enhancement image capable of reflecting the rut defects more accurately is determined, and further the defect detection result of the traffic road condition enhancement image is more accurate.
The method for acquiring the defect type in one embodiment of the invention comprises the following steps:
and carrying out defect detection on the traffic road condition enhanced image by using the CNN neural network, and determining whether defects exist and the defect type when the defects exist.
It should be noted that, the CNN neural network is a technical means well known to those skilled in the art, and is not described herein in detail, but only a brief process of determining defect types by using the CNN neural network in one embodiment of the present invention is described briefly:
And training the CNN neural network by using a reference data set, wherein the reference data set comprises a plurality of historical traffic road condition original images and the obtained traffic road condition enhanced images are obtained according to the mode. Because the part of traffic condition enhancement images comprise defect information, marking the part of traffic condition enhancement images according to defect conditions such as ruts, cracks, grooves and the like marks the traffic condition enhancement images with the same defect conditions such as ruts, cracks, grooves and the like as the same number, for example marking the part of traffic condition enhancement images without defect information as 0, marking the traffic condition enhancement images with ruts, cracks and grooves as 1, marking the traffic condition enhancement images with ruts, cracks and grooves as 2 and the like. Dividing the marked data set into a training set and a verification set according to a preset proportion, inputting a reference data set into a CNN neural network for training, wherein a loss function is cross entropy, and adopting a gradient descent method until the loss function converges to obtain the CNN neural network after training; and inputting the traffic road condition enhancement images into a CNN neural network after the training is completed, and automatically classifying the traffic road condition enhancement images into the defect category by the CNN neural network. In the embodiment of the invention, the preset ratio is 7: and 3, the implementer can set the implementation according to the implementation scene.
The invention also provides a traffic engineering road condition visual detection system, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor is used for running the corresponding computer program, and the computer program can realize the traffic engineering road condition visual detection method described in the steps when running in the processor.
In summary, the embodiment of the invention provides a traffic engineering road condition visual detection method and a traffic engineering road condition visual detection system, wherein firstly, in the embodiment of the invention, the distribution defect characteristic value of the gray level is obtained according to the distribution difference between the gray level and the number of pixels corresponding to the surrounding gray level and the proximity degree between the gray level and the gray level of a normal road surface; acquiring an imaging defect rule value of the gray level according to the correlation between all the clustering clusters corresponding to the gray level and the reference lane line connected domain and the distribution defect characteristic value of the gray level; further obtaining a defect probability value of the gray level; acquiring a traffic road condition enhancement image according to defect probability values of all gray levels of an original traffic road condition image; and finally, carrying out defect detection on the traffic road condition enhanced image. The invention improves the accuracy of detecting the visual defects of traffic engineering road conditions by effectively enhancing the prominence of the rut defects.
An embodiment of an image processing method for traffic engineering road condition visual detection is provided:
In order to enhance the defect expression level, the prior art uses a histogram equalization method to enhance the image, and the histogram equalization method redistributes the pixel values of the image to ensure that the brightness distribution of the image is more uniform, so that the contrast and detail of the defects in the image are enhanced, however, under the condition of no lower-angle light irradiation, the rut feature expression level is relatively weak, the histogram equalization method ignores the rut defect probability of the gray level, so that the contrast of the gray level with high defect probability and the gray level with low defect probability cannot be effectively amplified, and the salient level of the rut defect cannot be effectively enhanced.
In order to solve the technical problem, the embodiment provides an image processing method for traffic engineering road condition visual detection, which comprises the following steps:
step S1, an original image of the traffic road condition is obtained.
Step S2, obtaining a road condition gray level histogram of an original image of the traffic road condition; acquiring normal road surface gray levels according to the number of pixel points corresponding to all gray levels in the road condition gray level histogram; and in the road condition gray level histogram, acquiring a gray level distribution defect characteristic value according to the distribution difference between the gray level and the number of pixels corresponding to the surrounding gray level and the proximity degree between the gray level and the normal road surface gray level.
Step S3, binarizing the original image of the traffic road condition to obtain a reference lane line connected domain; clustering all pixel points corresponding to the gray level in an original image of the traffic road condition to obtain a plurality of clusters corresponding to the gray level; and acquiring an imaging defect rule value of the gray level according to the correlation between all the clustering clusters corresponding to the gray level and the reference lane line connected domain and the distribution defect characteristic value of the gray level.
Step S4, according to the distribution of the road condition gray level histogram, the imaging defect rule value of the gray level and the distribution defect characteristic value of the gray level, obtaining the defect probability value of the gray level; and carrying out enhancement processing on the original traffic road condition image according to the defect probability values of all gray levels of the original traffic road condition image to obtain an enhanced traffic road condition image.
Because the specific implementation process of steps S1 to S4 is already described in detail in the above-mentioned traffic engineering road condition visual detection method and system, the detailed description is omitted.
The beneficial effects of the embodiment of the invention include: according to the embodiment of the invention, the distribution defect characteristic value of the gray level is obtained according to the distribution difference between the gray level and the number of pixel points corresponding to the surrounding gray level and the proximity degree between the gray level and the gray level of a normal road surface; acquiring an imaging defect rule value of the gray level according to the correlation between all the clustering clusters corresponding to the gray level and the reference lane line connected domain and the distribution defect characteristic value of the gray level; further obtaining a defect probability value of the gray level; and obtaining the traffic road condition enhanced image according to the defect probability values of all gray levels of the traffic road condition original image. According to the defect probability value of the gray level, the invention effectively enhances the prominence degree of the rut defect.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (10)
1. The traffic engineering road condition visual detection method is characterized by comprising the following steps of:
Acquiring an original image of traffic road conditions;
Acquiring a road condition gray level histogram of the traffic road condition original image; acquiring normal road surface gray levels according to the number of pixel points corresponding to all gray levels in the road condition gray level histogram; in the road condition gray level histogram, according to the distribution difference between the gray level and the number of pixels corresponding to the surrounding gray level and the proximity degree between the gray level and the normal road surface gray level, obtaining a distribution defect characteristic value of the gray level;
Performing binarization processing on the traffic road condition original image to obtain a reference lane line connected domain; clustering all pixel points corresponding to the gray level in the traffic road condition original image to obtain a plurality of clustering clusters corresponding to the gray level; acquiring an imaging defect rule value of the gray level according to the correlation between all the clustering clusters corresponding to the gray level and the reference lane line connected domain and the distribution defect characteristic value of the gray level;
Acquiring a defect probability value of the gray level according to the distribution of the road condition gray level histogram, an imaging defect rule value of the gray level and a distribution defect characteristic value of the gray level; according to the defect probability values of all gray levels of the traffic road condition original image, enhancing the traffic road condition original image to obtain a traffic road condition enhanced image;
and carrying out defect detection on the traffic road condition enhanced image.
2. The traffic engineering road condition visual inspection method according to claim 1, wherein the method for obtaining the distribution defect characteristic value specifically comprises the following steps:
obtaining a variation characteristic parameter value according to a variation characteristic parameter value formula, wherein the variation characteristic parameter value formula comprises:
;/> in the road condition gray level histogram, the first/> Variable characteristic parameter values for the individual gray levels; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)In the road condition gray level histogram, the first/>The number of pixel points corresponding to the gray level; /(I)Is an absolute value symbol; /(I)A first adjustment factor that is denominator;
performing curve fitting according to the road condition gray level histogram to obtain a fitted gray level curve; segmenting a gray level sequence formed by gray levels of the road condition gray level histogram to obtain each gray level segment;
Obtaining a distribution defect characteristic value according to a distribution defect characteristic value formula, wherein the distribution defect characteristic value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)Is the gray level of a normal road surface; /(I)In the road condition gray level histogram, the first/>Variable characteristic parameter values for the individual gray levels; To at/> In the gray scale section to which each gray level belongs, all gray levels correspond to the variance of the derivative on the fitted gray scale curve; A second regulator that is denominator; /(I) In natural number/>Is an exponential function of the base.
3. The traffic engineering road condition visual inspection method according to claim 1, wherein the method for acquiring the imaging defect rule value comprises the following steps:
Acquiring a reference lane contour of a reference lane line connected domain; splitting the reference lane contour to obtain all the reference lane sectional contours of the reference lane contour; taking the longest reference lane segmentation contour as a reference driving contour edge; acquiring a reference driving characteristic sequence of the reference driving contour edge according to the distribution of pixel points in the communication domain of the reference driving contour edge and the reference lane line;
Obtaining a cluster contour and a cluster area of each cluster corresponding to each gray level by using a convex hull detection method; splitting each class cluster contour to obtain all class cluster segmentation contours of each class cluster contour; acquiring a running characteristic sequence to be analyzed of each cluster segmentation contour according to each cluster segmentation contour and the corresponding pixel point distribution in the cluster region;
And acquiring imaging defect rule values of gray levels according to the correlation between the reference driving characteristic sequence and the driving characteristic sequence to be analyzed of each cluster sectional profile, the distribution defect characteristic values of gray levels and the fluctuation of the gray levels corresponding to the lengths of all cluster sectional profiles.
4. The traffic engineering road condition visual inspection method according to claim 3, wherein the method for obtaining the imaging defect rule value comprises the following steps:
acquiring an imaging defect rule value according to an imaging defect rule value formula, wherein the imaging defect rule value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>Imaging defect rule values for the individual gray levels; /(I)For/>The number of gray levels corresponds to the total number of all class cluster segmented contours; /(I)For/>The gray level corresponds to the/>The length of the individual cluster segmentation profile; /(I)For/>The gray levels correspond to the average value of the lengths of the segmented contours of all the class clusters; /(I)In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)For/>The gray level corresponds to the/>A running characteristic sequence to be analyzed of the sectional profile of each cluster; /(I)A reference travel feature sequence that is a reference travel profile edge; for/> The individual gray levels correspond to/> of the segmented contours of all class clustersMaximum value of (2); /(I)Is an absolute value symbol; /(I)To take the pearson correlation coefficient function.
5. The traffic engineering road condition visual inspection method according to claim 3, wherein the method for acquiring the reference driving characteristic sequence and the driving characteristic sequence to be analyzed comprises the following steps:
Taking any one of the reference driving contour edges or the cluster-like segmented contours as a target analysis edge; taking any cluster area or a reference lane line connected area as a target analysis area; taking all the pixel points in the target analysis area as the pixel points of the target area; taking all pixel points on the target analysis edge as target edge pixel points;
in the target analysis area to which the target analysis edge belongs, sequentially counting the total number of all target area pixel points of the target edge pixel points in a preset direction by taking the end point of any one of the target analysis edges as a starting point and the extending direction of the target analysis edge as a direction, so as to obtain a target running characteristic sequence of the target analysis edge;
And taking the target running characteristic sequence of the reference running contour edge as the reference running characteristic sequence of the reference running contour edge, and taking the target running characteristic sequence of the cluster-like segmented contour as the running characteristic sequence to be analyzed of the cluster-like segmented contour.
6. The traffic engineering road condition visual inspection method according to claim 1, wherein the defect probability value obtaining method comprises:
obtaining a defect probability value according to a defect probability value formula, wherein the defect probability value formula comprises:
; wherein/> In the road condition gray level histogram, the first/>-Said defect probability values for individual gray levels; /(I)In the road condition gray level histogram, the first/>Imaging defect rule values for the individual gray levels; /(I)In the road condition gray level histogram, the first/>Distribution defect feature values of the individual gray levels; /(I)The deviation of the road condition gray level histogram is obtained; /(I)The kurtosis of the road condition gray level histogram; /(I)In natural number/>Is an exponential function of the base.
7. The traffic engineering road condition visual inspection method according to claim 1, wherein the method for acquiring the normal road surface gray level comprises the following steps:
And taking the gray level corresponding to the maximum number of pixels in the road condition gray level histogram as the normal road surface gray level.
8. The traffic engineering road condition visual inspection method according to claim 1, wherein the method for obtaining the reference lane line connected domain specifically comprises:
Based on an Ojin threshold method, carrying out binarization processing on the traffic road condition original image to obtain a traffic binarization image; and extracting each front Jing Liantong domain of the traffic binarization image by using a connected domain analysis algorithm, and taking the largest front Jing Liantong domain as a reference lane line connected domain.
9. The traffic engineering road condition visual inspection method according to claim 1, wherein the method for acquiring the traffic road condition enhanced image comprises the following steps:
Acquiring an initial mapped gray level of the gray level based on a histogram equalization method;
Acquiring updated mapped gray levels of each gray level according to the defect probability value corresponding to the gray level and the initial mapped gray level of the gray level; the defect probability value and the updated mapped gray level show positive correlation; the initial mapped gray level and the updated mapped gray level exhibit a positive correlation;
And based on a histogram equalization method, acquiring a traffic road condition enhanced image according to the updated mapped gray level of each gray level.
10. A traffic engineering road condition visual detection system comprising a memory, a processor and a computer program stored in the memory and operable on the processor, wherein the processor, when executing the computer program, implements the steps of a traffic engineering road condition visual detection method as claimed in any one of claims 1 to 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410425015.6A CN118015002B (en) | 2024-04-10 | 2024-04-10 | Traffic engineering road condition visual detection method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410425015.6A CN118015002B (en) | 2024-04-10 | 2024-04-10 | Traffic engineering road condition visual detection method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN118015002A true CN118015002A (en) | 2024-05-10 |
CN118015002B CN118015002B (en) | 2024-06-18 |
Family
ID=90952400
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410425015.6A Active CN118015002B (en) | 2024-04-10 | 2024-04-10 | Traffic engineering road condition visual detection method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN118015002B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118196092A (en) * | 2024-05-17 | 2024-06-14 | 智汇新能科技(大连)有限公司 | Automatic detection method and system for highway construction quality |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103903268A (en) * | 2014-04-09 | 2014-07-02 | 南京理工大学 | Road surface crack detection method based on multiple lower-layer marked features |
CN104792792A (en) * | 2015-04-27 | 2015-07-22 | 武汉武大卓越科技有限责任公司 | Stepwise-refinement pavement crack detection method |
CN107945161A (en) * | 2017-11-21 | 2018-04-20 | 重庆交通大学 | Road surface defect inspection method based on texture feature extraction |
US20180137612A1 (en) * | 2015-04-27 | 2018-05-17 | Wuhan Wuda Zoyon Science And Technology Co., Ltd. | A stepwise refinement detection method for pavement cracks |
CN115457041A (en) * | 2022-11-14 | 2022-12-09 | 安徽乾劲企业管理有限公司 | Road quality identification and detection method |
CN116934748A (en) * | 2023-09-15 | 2023-10-24 | 山东重交路桥工程有限公司 | Pavement crack detection system based on emulsified high-viscosity asphalt |
CN117635611A (en) * | 2024-01-25 | 2024-03-01 | 大连傲盈科技有限公司 | Traffic construction pavement quality monitoring method and system |
-
2024
- 2024-04-10 CN CN202410425015.6A patent/CN118015002B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103903268A (en) * | 2014-04-09 | 2014-07-02 | 南京理工大学 | Road surface crack detection method based on multiple lower-layer marked features |
CN104792792A (en) * | 2015-04-27 | 2015-07-22 | 武汉武大卓越科技有限责任公司 | Stepwise-refinement pavement crack detection method |
US20180137612A1 (en) * | 2015-04-27 | 2018-05-17 | Wuhan Wuda Zoyon Science And Technology Co., Ltd. | A stepwise refinement detection method for pavement cracks |
CN107945161A (en) * | 2017-11-21 | 2018-04-20 | 重庆交通大学 | Road surface defect inspection method based on texture feature extraction |
CN115457041A (en) * | 2022-11-14 | 2022-12-09 | 安徽乾劲企业管理有限公司 | Road quality identification and detection method |
CN116934748A (en) * | 2023-09-15 | 2023-10-24 | 山东重交路桥工程有限公司 | Pavement crack detection system based on emulsified high-viscosity asphalt |
CN117635611A (en) * | 2024-01-25 | 2024-03-01 | 大连傲盈科技有限公司 | Traffic construction pavement quality monitoring method and system |
Non-Patent Citations (1)
Title |
---|
CIPRIAN SORANDARU ET AL.: "Detecting Road Irregularities by Image Enhancements and Thresholding", 《IEEE EUROCON 2019 -18TH INTERNATIONAL CONFERENCE ON SMART TECHNOLOGIES》, 11 October 2019 (2019-10-11), pages 1 - 6 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118196092A (en) * | 2024-05-17 | 2024-06-14 | 智汇新能科技(大连)有限公司 | Automatic detection method and system for highway construction quality |
Also Published As
Publication number | Publication date |
---|---|
CN118015002B (en) | 2024-06-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116758059B (en) | Visual nondestructive testing method for roadbed and pavement | |
CN118015002B (en) | Traffic engineering road condition visual detection method and system | |
CN115311292B (en) | Strip steel surface defect detection method and system based on image processing | |
CN103400151B (en) | The optical remote sensing image of integration and GIS autoregistration and Clean water withdraw method | |
CN115908411B (en) | Concrete curing quality analysis method based on visual detection | |
CN112651968B (en) | Wood board deformation and pit detection method based on depth information | |
CN111191628B (en) | Remote sensing image earthquake damage building identification method based on decision tree and feature optimization | |
CN110263635B (en) | Marker detection and identification method based on structural forest and PCANet | |
CN115115641B (en) | Pupil image segmentation method | |
CN116385450B (en) | PS sheet wear resistance detection method based on image processing | |
CN108921076B (en) | Pavement crack disease self-adaptive constant false alarm detection method based on image | |
CN107705294B (en) | Cross laser image type roadbed surface settlement monitoring method and monitoring system | |
CN116630813B (en) | Highway road surface construction quality intelligent detection system | |
CN113239733B (en) | Multi-lane line detection method | |
CN117173661B (en) | Asphalt road quality detection method based on computer vision | |
CN116342586B (en) | Road surface quality detection method based on machine vision | |
CN115909256B (en) | Road disease detection method based on road visual image | |
CN114820625A (en) | Automobile top block defect detection method | |
CN114894804A (en) | Method for detecting surface cracks of precision standard part | |
CN110473224B (en) | Automatic RSF level set image segmentation method based on KL entropy | |
CN106803066B (en) | Vehicle yaw angle determination method based on Hough transformation | |
CN116258864B (en) | Village planning construction big data management system | |
CN115797374B (en) | Airport runway extraction method based on image processing | |
CN114677428A (en) | Power transmission line icing thickness detection method based on unmanned aerial vehicle image processing | |
CN114419317A (en) | Light strip center extraction method for light with complex environment line structure |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |