CN117237245B - Industrial material quality monitoring method based on artificial intelligence and Internet of things - Google Patents
Industrial material quality monitoring method based on artificial intelligence and Internet of things Download PDFInfo
- Publication number
- CN117237245B CN117237245B CN202311525547.9A CN202311525547A CN117237245B CN 117237245 B CN117237245 B CN 117237245B CN 202311525547 A CN202311525547 A CN 202311525547A CN 117237245 B CN117237245 B CN 117237245B
- Authority
- CN
- China
- Prior art keywords
- gray
- allocated
- value
- gray value
- values
- 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.)
- Active
Links
- 239000012770 industrial material Substances 0.000 title claims abstract description 87
- 238000000034 method Methods 0.000 title claims abstract description 70
- 238000012544 monitoring process Methods 0.000 title claims abstract description 32
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 20
- 238000001914 filtration Methods 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 16
- 230000007547 defect Effects 0.000 claims description 11
- 238000012216 screening Methods 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 4
- 235000002918 Fraxinus excelsior Nutrition 0.000 claims 1
- 239000002956 ash Substances 0.000 claims 1
- 238000009826 distribution Methods 0.000 abstract description 17
- 230000000694 effects Effects 0.000 abstract description 4
- 238000012545 processing Methods 0.000 abstract description 4
- 230000009467 reduction Effects 0.000 abstract description 4
- 239000000463 material Substances 0.000 description 14
- 239000002245 particle Substances 0.000 description 14
- 238000012549 training Methods 0.000 description 5
- 238000011946 reduction process Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Image Processing (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to an industrial material quality monitoring method based on artificial intelligence and the Internet of things, which comprises the following steps: collecting an industrial material gray level image of the industrial material; dividing an industrial material gray level image into a plurality of image blocks, and obtaining the gray level connected domain duty ratio according to the image blocks; obtaining the noise-containing degree according to the gray scale connected domain duty ratio; obtaining a main body gray scale interval formed by gray values to be distributed from all gray values in an image block; obtaining a pixel point distribution range and the number of pixel points to be distributed according to the noise-containing degree; obtaining a contrast gray level histogram according to the number of pixel points to be distributed; obtaining contrast difference according to the contrast gray level histogram; obtaining an ideal gray value according to the contrast difference; and carrying out quality monitoring on the industrial gray denoising image according to the ideal gray value. The invention has better noise reduction effect and improves the accuracy of the quality monitoring result of the industrial materials.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an industrial material quality monitoring method based on artificial intelligence and the Internet of things.
Background
The quality detection of industrial materials is an important link in the industrial production process, and mainly through carrying out automatic quality monitoring on industrial material images, certain noise exists in the industrial material images due to the influence of acquisition equipment and environment, and in order to ensure the accuracy of quality monitoring, the industrial material images need to be denoised.
The traditional method generally adopts a wiener filtering algorithm to denoise the industrial material image, but because the material particles in the industrial material image are similar in size, shape and color and are distributed in a scattered manner, the noise and the material particles are difficult to distinguish, so that the traditional wiener filtering can excessively reduce noise or reduce noise in the noise reduction process.
Disclosure of Invention
The invention provides an industrial material quality monitoring method based on artificial intelligence and the Internet of things, which aims to solve the existing problems: because the size, shape and color of the material particles in the industrial material image are similar, and the material particles are distributed in a scattered manner, the noise and the material particles are difficult to distinguish, and the traditional wiener filtering can excessively reduce noise or reduce noise in the noise reduction process.
The invention discloses an industrial material quality monitoring method based on artificial intelligence and the Internet of things, which adopts the following technical scheme:
the method comprises the following steps:
collecting industrial material gray level images of a plurality of industrial materials;
dividing an industrial material gray level image into a plurality of image blocks, and counting the number of a plurality of connected domains of each gray level value in the image blocks to obtain the gray level connected domain duty ratio of each gray level value; obtaining the noise-containing degree of each gray value according to the gray connected domain duty ratio, wherein the noise-containing degree is used for describing the difference of the gray value compared with the noise contained in other gray values;
screening a plurality of main body gray scale intervals formed by gray scale values to be distributed from intervals formed by all gray scale values in the image block; carrying out gray value range judgment according to the noise degree of the gray values to be allocated and the number of the gray values to be allocated in the main gray interval to obtain the pixel point allocation range of each gray value to be allocated; according to the noise degree of the gray values to be allocated and the number of the pixel points of the gray values to be allocated, judging the number of the pixel points to be allocated to obtain the number of the pixel points to be allocated of each gray value to be allocated; gray scale adjustment is carried out on the pixel points of the gray values to be allocated according to the number of the pixel points to be allocated to obtain a contrast gray scale histogram of each gray value to be allocated; obtaining contrast difference of each gray value to be allocated according to the difference between the gray value corresponding pixel points in the contrast gray histogram, wherein the contrast difference is used for describing the difference between the image blocks of the industrial material gray image and the image blocks of the industrial material gray image without noise interference;
screening ideal gray values from a plurality of gray values to be allocated according to the contrast difference; and carrying out defect monitoring on the industrial gray denoising image according to the ideal gray value.
Preferably, the counting of the number of the connected domains of each gray value in the image block to obtain the gray connected domain duty ratio of each gray value includes the following specific steps:
for any image block, carrying out connected domain analysis on the image block to obtain a plurality of connected domains of the image block, and obtaining a gray level histogram of the image block; for any gray value in the gray histogram of the image block, each connected domain of the gray value is marked as one gray connected domain of the gray value, all the gray connected domains of all the gray values are obtained, the number of the gray connected domains of all the gray values is subjected to linear normalization, and each normalized number is marked as the gray connected domain duty ratio of the gray value.
Preferably, the noise level of each gray value is obtained according to the gray connected domain duty ratio, which comprises the following specific steps:
for any image block, carrying out connected domain analysis on the image block to obtain a plurality of connected domains of the image block, and obtaining a gray level histogram of the image block; for any gray value in the gray histogram of the image block, marking each connected domain of the gray value as one gray connected domain of the gray value;
for any gray connected domain, acquiring gradient amplitude values of all pixel points in the gray connected domain by using a sobel operator, and marking the average value of the gradient amplitude values of all pixel points in the gray connected domain as the gray gradient amplitude value of the gray connected domain;
in the method, in the process of the invention,the noise level of the gray value is represented; />Gray representing gray valueA degree connected domain duty cycle; />Representing the variance of the gray gradient amplitude values of all gray connected domains under the gray value; />An exponential function based on a natural constant is represented.
Preferably, the method for screening out a plurality of main gray scale intervals composed of gray scale values to be allocated from intervals composed of all gray scale values in the image block includes the following specific steps:
the preset number of gray values is marked as T2, for any image block, the gray value with the largest number of pixels in the gray histogram of the image block is marked as a main body center gray value, a section formed by the gray values with the number of continuous T2 pixels which are not 0 before the main body center gray value and the gray values with the number of continuous T2 pixels which are not 0 after the main body center gray value is marked as a main body gray section of the image block, and each gray value in the main body gray section is marked as a gray value to be allocated.
Preferably, the gray value range determination is performed according to the noise level of the gray value to be allocated and the number of the gray values to be allocated in the main gray interval, so as to obtain the pixel point allocation range of each gray value to be allocated, and the specific method includes:
for any one gray value to be allocated in the main gray interval of any one image block, in the formula,a pixel point allocation range representing a gray value to be allocated; />Representing the noise level of the gray value to be allocated; />Representing the number of all gray values to be allocated in the main gray interval; />Representing an upward rounding.
Preferably, the determining the number of the pixel points according to the noise level of the gray value to be allocated and the number of the pixel points of the gray value to be allocated to obtain the number of the pixel points to be allocated of each gray value to be allocated includes the following specific steps:
for any one gray value to be allocated in the main gray interval of any one image block, in the formula,representing the number of pixel points to be allocated of the gray value to be allocated; />Representing the noise level of the gray value to be allocated; />Representing the gray value to be allocated with the largest pixel number; />Representing a gray value to be assigned; />Representing the number of all gray values to be allocated in the main gray interval; />Representing the number of pixels to be assigned with gray values; />The representation takes absolute value; />Representing an upward rounding.
Preferably, the gray adjustment is performed on the pixel points of the gray values to be allocated according to the number of the pixel points to be allocated to obtain a contrast gray histogram of each gray value to be allocated, which comprises the following specific steps:
for any gray value to be allocated in the main gray interval of any image block, the gray value to be allocated is precededAfter the gray value and the gray value to be allocated +.>The gray values are all marked as gray values to be adjusted of the gray values to be allocated; randomly acquiring the number of pixel points to be allocated in the gray value to be allocated, and marking the pixel points to be allocated as the gray value to be allocated;
performing a gray scale adjustment operation, the gray scale adjustment operation comprising: the gray value of any pixel point to be allocated with the gray value which is not adjusted is adjusted to be the first gray value to be adjusted by taking the first gray value to be adjusted of the gray value to be allocated as a starting point; the gray value of any pixel point to be allocated which is not adjusted with the gray value is adjusted to be a second gray value to be adjusted; adjusting the gray value of any pixel point to be allocated with the gray value which is not adjusted to be a third gray value to be adjusted, and the like until the pixel point to be allocated with the gray value which is not adjusted is not existed or all the gray values to be adjusted are traversed;
if all the gray values to be adjusted are traversed, pixel points to be allocated with unadjusted gray values exist, and the gray adjustment operation is repeated until no pixel points to be allocated with unadjusted gray values exist;
and marking the corresponding gray level histogram after gray level adjustment of all the allocated pixel points of the gray level value to be allocated as a contrast gray level histogram of the gray level value to be allocated.
Preferably, the obtaining the contrast difference of each gray value to be allocated according to the difference between the gray value corresponding to the pixel point number in the contrast gray histogram includes the following specific methods:
for any one gray value to be allocated in the main gray interval of any one image block, in the formula,representing a contrast difference in gray values to be assigned; />The deviation of the number of the pixel points of all the gray values to be allocated in the main gray interval in the contrast gray histogram of the gray values to be allocated; />And the overstep kurtosis of the number of the pixel points of all the gray values to be allocated in the main gray interval in the contrast gray histogram of the gray values to be allocated.
Preferably, the method for selecting the ideal gray value from the plurality of gray values to be allocated according to the contrast difference comprises the following specific steps:
and for any image block, marking the gray value to be allocated with the largest contrast difference as an ideal gray value of the image block.
Preferably, the defect monitoring of the industrial gray-scale denoising image according to the ideal gray-scale value comprises the following specific steps:
for any one image block in any one industrial material gray level image, in a main gray level interval on a gray level histogram of the image block, marking the serial number of each gray level value to be allocated of the image block as the iteration times of the image block; the serial number of the ideal gray value of the image block is marked as the ideal iteration number of the image block;
in the method, in the process of the invention,filter regulation parameters representing image blocks: />Representing an ideal iteration number of the image block; />Representing the maximum value of all iteration times of the image block;
replacing smoothing parameters of wiener filtering with filtering regulation parameters of the image blocks, carrying out wiener filtering on the image blocks to obtain denoised image blocks, and marking the industrial material gray level image as an industrial material gray level denoising image after all denoised image blocks in the industrial material gray level image are obtained;
inputting the industrial material gray level denoising image into a trained neural network to obtain a plurality of defect areas in the industrial material gray level denoising image, and obtaining a plurality of defect areas of all the industrial gray level denoising images.
The technical scheme of the invention has the beneficial effects that: obtaining the noise degree of the gray value according to the gray image of the industrial material, and judging the number according to the noise degree to obtain the pixel distribution range of the gray value to be distributed and the number of the pixels to be distributed; obtaining a contrast gray level histogram according to the pixel point distribution range and the number of the pixel points to be distributed, obtaining contrast difference of gray level values to be distributed according to the contrast gray level histogram, obtaining ideal gray level values according to the contrast difference, and performing quality monitoring on the industrial gray level denoising image according to the ideal gray level values; compared with the prior art, the size, shape and color of the material particles in the industrial material image are similar, and the material particles are distributed in a scattered manner, so that the noise and the material particles are difficult to distinguish, and the situation that the traditional wiener filtering can excessively reduce noise or reduce noise in the noise reduction process is caused; the noise-containing degree of the invention reflects the difference of the gray value compared with the noise contained in other gray values, and the contrast difference reflects the difference between the image block of the gray image of the industrial material and the image block of the gray image of the industrial material without noise interference; the probability of excessive noise reduction and insufficient noise reduction is reduced, so that the noise reduction effect is better, and the accuracy of the industrial material quality monitoring result is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of steps of an industrial material quality monitoring method based on artificial intelligence and the Internet of things.
Detailed Description
In order to further illustrate the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of the specific implementation, structure, characteristics and effects of the industrial material quality monitoring method based on artificial intelligence and the internet of things according to the invention, with reference to the accompanying drawings and the preferred embodiment. 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 specific scheme of an industrial material quality monitoring method based on artificial intelligence and the Internet of things, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of steps of an industrial material quality monitoring method based on artificial intelligence and internet of things according to an embodiment of the present invention is shown, where the method includes the following steps:
step S001: and collecting industrial material gray level images of a plurality of industrial materials.
It should be noted that, in the conventional method, the wiener filtering algorithm is generally adopted to denoise the industrial material image, but because the material particles in the industrial material image are similar in size, shape and color, and the material particles are distributed in a scattered manner, the noise and the material particles are difficult to distinguish, so that the conventional wiener filtering can excessively reduce noise or reduce noise insufficiently in the noise reduction process. Therefore, the embodiment provides an industrial material quality monitoring method based on artificial intelligence and the Internet of things.
Specifically, in order to implement the industrial material quality monitoring method based on the artificial intelligence and the internet of things provided by the embodiment, firstly, an industrial material gray level image needs to be acquired, and the specific process is as follows: and shooting industrial material images of a plurality of industrial materials by using an industrial camera, and carrying out graying treatment on each industrial material image to obtain a plurality of industrial material gray images. The graying process is a known technique, and the description of this embodiment is omitted.
So far, the industrial material gray level images of a plurality of industrial materials are obtained through the method.
Step S002: dividing an industrial material gray level image into a plurality of image blocks, and counting the number of a plurality of connected domains of each gray level value in the image blocks to obtain the gray level connected domain duty ratio of each gray level value; and obtaining the noise degree of each gray value according to the gray connected domain duty ratio.
It should be noted that, the industrial material gray level image contains a large amount of material particles, the size and shape of the material particles are similar, the colors are similar, the distribution condition of the gray value corresponding to the number of the pixel points in the ideal gray level histogram shows that the middle area is high, the two end areas are low, and the distribution is similar to Gaussian distribution; since noise is generally randomly and discretely distributed and the overall number is small, there is a large difference between different noise points; for the gray level histogram interfered by noise, the noise can interfere the distribution condition of the gray level value corresponding to the number of the pixel points in the gray level histogram, so that the gray level value is less approximate to Gaussian distribution; therefore, the embodiment can obtain the noise-containing degree of each gray value by analyzing the distribution condition of the corresponding connected domain under different gray values, and determine the difference between the gray image of the industrial material and the gray image of the industrial material without noise interference through the noise-containing degree, thereby determining the noise intensity in the gray image of the industrial material so as to facilitate the subsequent analysis and treatment.
Specifically, a block number T1 is preset, where the present embodiment is described by taking t1=16 as an example, and the present embodiment is not specifically limited, where T1 may be determined according to specific implementation situations; taking any industrial material gray level image as an example, dividing the industrial material gray level image into T1 image blocks uniformly; taking any image block as an example, carrying out connected domain analysis on the image block to obtain a plurality of connected domains of the image block, and obtaining a gray level histogram of the image block; taking any gray value in the gray histogram of the image block as an example, marking each connected domain of the gray value as one gray connected domain of the gray value, acquiring all gray connected domains of all gray values, carrying out linear normalization on the number of the gray connected domains of all gray values, and marking each normalized number as the gray connected domain duty ratio of the gray value. Wherein each connected domain corresponds to a gray value, and one gray value may correspond to a plurality of connected domains; the abscissa in the gray level histogram represents the gray level value, the ordinate represents the number of pixels corresponding to the gray level value, and the acquisition of the gray level histogram is a known technique, which is not described in detail in this embodiment.
Further, taking any one gray scale connected domain of any one gray scale value in the gray scale histogram of the image block as an example, acquiring gradient amplitudes of all pixel points in the gray scale connected domain by using a sobel operator, and marking the average value of the gradient amplitudes of all pixel points in the gray scale connected domain as the gray scale gradient amplitude of the gray scale connected domain; and acquiring the gray gradient amplitude value of each gray connected domain of the gray value. And obtaining the noise-containing degree of the gray value according to the gray connected domain duty ratio of the gray value and the gray gradient amplitude of each gray connected domain under the gray value. The sobel operator is a known technique, and this embodiment will not be described in detail. In addition, the method for calculating the noise-containing degree of the gray value comprises the following steps:
in the method, in the process of the invention,a noise level representing the gray value; />A gradation connected domain duty ratio representing the gradation value; />Representing the variance of the gray gradient amplitude values of all gray connected domains under the gray value; />An exponential function that is based on a natural constant; example use->The functions are presented with inverse proportion relation and normalization processing, and an implementer can select the inverse proportion function and the normalization function according to actual conditions. The larger the noise-containing degree of the gray value is, the larger the difference between the gray connected domains under the gray value is, and the more scattered the gray connected domains under the gray value are distributed in the image block, the greater the possibility of reflecting that the gray value has noise is. Acquiring the noise level of all gray values in a gray histogram of the image block; and obtaining the noise degree of all gray values in the gray histograms of all image blocks.
So far, the noise-containing degree of all gray values in the gray level histograms of all the image blocks is obtained through the method.
Step S003: screening a plurality of main body gray scale intervals formed by gray scale values to be distributed from intervals formed by all gray scale values in the image block; carrying out gray value range judgment according to the noise degree of the gray values to be allocated and the number of the gray values to be allocated in the main gray interval to obtain the pixel point allocation range of each gray value to be allocated; according to the noise degree of the gray values to be allocated and the number of the pixel points of the gray values to be allocated, judging the number of the pixel points to be allocated to obtain the number of the pixel points to be allocated of each gray value to be allocated; gray scale adjustment is carried out on the pixel points of the gray values to be allocated according to the number of the pixel points to be allocated to obtain a contrast gray scale histogram of each gray value to be allocated; and obtaining the contrast difference of each gray value to be distributed according to the difference between the gray values in the contrast gray histogram and the number of the corresponding pixel points.
In the process of determining the noise intensity in the gray level image of the industrial material, the color of the material particles contained in the gray level image of the industrial material is similar, so that a gray value area with a certain continuous range exists in the corresponding gray level histogram to correspond to the gray value area; the plurality of gray values contained in this gray value region have a large influence on the quality monitoring result. For any gray value in the gray value region, the larger the noise content of the gray value is, the higher the noise content of the gray value is, the larger the interference to the gaussian distribution with ideal pixel number is, and therefore the larger the degree of adjustment is needed; according to the embodiment, the original distribution characteristics can be changed by adjusting the number of the pixels corresponding to the gray value according to the noise level, so that the subsequent analysis and processing can be facilitated.
Specifically, a number of gray values T2 is preset, where the present embodiment is described by taking t2=20 as an example, and the present embodiment is not specifically limited, where T2 may be determined according to specific implementation situations; taking any image block as an example, marking a gray value with the largest number of pixels in a gray histogram of the image block as a main body center gray value, marking a section formed by continuously taking a gray value with the number of T2 pixels which is not 0 before the main body center gray value and a gray value with the number of T2 pixels which is not 0 after the main body center gray value as a main body gray section of the image block, and marking each gray value in the main body gray section as a gray value to be allocated. In addition, when the number of gray values, which are not 0, of consecutive pixels before and after the gray value of the center of the main body does not satisfy T2, the main body gray interval is acquired based on the number of gray values that are actually present. Wherein each image block corresponds to a subject gray scale interval.
Further, taking any gray value to be allocated in the main gray interval as an example, the pixel allocation range of the gray value to be allocated is obtained according to the noise level of the gray value to be allocated. The method for calculating the pixel point distribution range of the gray value to be distributed comprises the following steps:
in the method, in the process of the invention,a pixel point distribution range for representing the gray value to be distributed; />Representing the noise level of the gray value to be allocated; />Representing the number of all gray values to be allocated in the main gray interval; />Representing an upward rounding. If the distribution range of the pixel points with gray values to be distributed is larger, the larger the distribution range of the pixel points with gray values to be distributed is, the larger the adjustable gray value range of the pixel points corresponding to the gray values to be distributed is, and the larger the interference of the original number of the pixel points with gray values to be distributed to the ideal number of the pixel points is reflected.
Further, the number of pixels to be allocated with the gray value to be allocated is obtained according to the noise level of the gray value to be allocated and the number of pixels with the gray value to be allocated. The method for calculating the number of the pixel points to be allocated of the gray value to be allocated comprises the following steps:
in the method, in the process of the invention,representing the number of pixel points to be allocated of the gray value to be allocated; />Representing the gray value to be allocatedIs a noise level of (1); />Representing the gray value to be allocated with the largest pixel number; />Representing the gray value to be allocated; />Representing the number of all gray values to be allocated in the main gray interval; />Representing the number of pixel points of the gray value to be allocated; />The representation takes absolute value; />Representing an upward rounding. If the number of the pixels to be allocated with the gray value to be allocated is larger, the difference between the original number of the pixels to be allocated with the gray value to be allocated and the number of the pixels to be allocated with the surrounding gray value to be allocated is larger, and the degree of the gray value to be allocated to be adjusted is reflected to be larger.
Further, before the gray value to be allocatedGray values and the gray value to be assigned +.>The gray values are all marked as the gray values to be adjusted of the gray values to be allocated; randomly acquiring the number of pixels to be allocated from the gray value to be allocated, and marking the pixels as the pixels to be allocated of the gray value to be allocated; and carrying out gray adjustment on all the allocated pixel points of the gray value to be allocated, wherein the specific process is as follows:
the gray value of any pixel point to be allocated with the gray value which is not adjusted is adjusted to be the first gray value by taking the first gray value to be adjusted of the gray value to be allocated as the starting pointAdjusting the gray value; the gray value of any pixel point to be allocated which is not adjusted with the gray value is adjusted to be a second gray value to be adjusted; adjusting the gray value of any pixel point to be allocated with the gray value which is not adjusted to be a third gray value to be adjusted, and the like until the pixel point to be allocated with the gray value which is not adjusted is not existed or all the gray values to be adjusted are traversed; if all the gray values to be adjusted are traversed, the pixel points to be allocated with the gray values which are not adjusted are still existed, and the operation is repeated until the pixel points to be allocated with the gray values which are not adjusted are not existed. It should be noted that if the number of gray values actually existing before the gray value to be assigned is not satisfiedAfter the gray value to be allocated is +.>The gray values are all marked as the gray values to be adjusted of the gray values to be allocated; if the number of grey values actually present after the grey value to be assigned does not satisfy +.>Before the gray value to be assigned +.>The gray values are all marked as the gray values to be adjusted of the gray values to be allocated.
Further, the gray level histogram corresponding to the gray level adjustment of all the allocated pixel points of the gray level value to be allocated is recorded as the contrast gray level histogram of the gray level value to be allocated. In addition, the gray value contained in the main gray interval of the gray histogram of the image block is consistent with the gray value contained in the main gray interval of the contrast gray histogram of the gray value to be allocated.
And obtaining the contrast difference of the gray value to be allocated according to the number of the pixel points of the gray value to be allocated in the contrast gray histogram. The calculation method of the contrast difference of the gray values to be distributed comprises the following steps:
in the method, in the process of the invention,representing a contrast difference of the gray value to be assigned; />Representing the deviation of the number of pixel points of all the gray values to be allocated in a main gray interval in the contrast gray histogram of the gray values to be allocated; />And the overstep kurtosis of the number of the pixel points of all the gray values to be allocated in the main gray interval in the contrast gray histogram of the gray values to be allocated is represented. If the contrast difference of the gray value to be allocated is smaller, it is indicated that after all the allocated pixels of the gray value to be allocated are adjusted, the number distribution of the pixels in the corresponding histogram tends to be ideal gaussian distribution, and the interference of the gray value to be allocated to the detection result is reflected to be larger. Acquiring contrast differences of all gray values to be allocated of the image block; acquiring contrast differences of all gray values to be allocated in the image block, and marking the gray value to be allocated with the largest contrast difference as an ideal gray value of the image block; and acquiring ideal gray values of all the image blocks. The obtaining of the skewness and the over-peak kurtosis is a known technique, and the embodiment is not repeated here.
So far, the ideal gray values of all the image blocks are obtained through the method.
Step S004: screening ideal gray values from a plurality of gray values to be allocated according to the contrast difference; and carrying out defect monitoring on the industrial gray denoising image according to the ideal gray value.
Specifically, taking any image block as an example, in a main gray scale interval on a gray scale histogram of the image block, marking a serial number of each gray scale value to be allocated of the image block as the iteration number of the image block; the serial number of the ideal gray value of the image block is marked as the ideal iteration number of the image block; and obtaining the filtering regulation parameters of the image block according to the number of iterations and the ideal number of iterations of the image block. The method for calculating the filtering regulation parameters of the image block comprises the following steps:
in the method, in the process of the invention,filter control parameters representing the image block: />Representing an ideal number of iterations for the image block; />Representing the maximum of all iterations of the image block. The larger the filtering regulation parameters of the image block, the larger the difference between the image block and an ideal noiseless image block is, the larger the noise intensity of the image block is, and the larger the degree of the image block to be regulated is reflected.
Further, the smoothing parameters of wiener filtering are replaced by the filtering regulation parameters of the image block, the image block is subjected to wiener filtering to obtain a denoised image block, all denoised image blocks in the industrial material gray level image are obtained, and the industrial material gray level image at the moment is recorded as an industrial material gray level denoised image. The denoising with smoothing parameters is a well known content of wiener filtering, and this embodiment is not repeated here.
Further, inputting the industrial material gray level denoising image into a trained neural network to obtain a plurality of defect areas in the industrial material gray level denoising image, and obtaining a plurality of defect areas of all industrial gray level denoising images; the neural network used in this embodiment is YOLOv3, and the method for acquiring the data set for training the neural network is as follows: collecting a large number of industrial material gray level denoising images, artificially marking a bounding box in each industrial material gray level denoising image, and marking the marking result as a label of each industrial material gray level denoising image; collecting a large number of industrial material gray level denoising images and corresponding labels thereof to form a data set; training the neural network by using the data set, wherein a loss function used in the training process is a mean square error loss function; the specific training process is a well-known content of the neural network, and the specific training process is not described in detail in this embodiment.
This embodiment is completed.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the invention, but any modifications, equivalent substitutions, improvements, etc. within the principles of the present invention should be included in the scope of the present invention.
Claims (8)
1. An industrial material quality monitoring method based on artificial intelligence and the Internet of things is characterized by comprising the following steps:
collecting industrial material gray level images of a plurality of industrial materials;
dividing an industrial material gray level image into a plurality of image blocks, and counting the number of a plurality of connected domains of each gray level value in the image blocks to obtain the gray level connected domain duty ratio of each gray level value; obtaining the noise-containing degree of each gray value according to the gray connected domain duty ratio, wherein the noise-containing degree is used for describing the difference of the gray value compared with the noise contained in other gray values;
screening a plurality of main body gray scale intervals formed by gray scale values to be distributed from intervals formed by all gray scale values in the image block; carrying out gray value range judgment according to the noise degree of the gray values to be allocated and the number of the gray values to be allocated in the main gray interval to obtain the pixel point allocation range of each gray value to be allocated; according to the noise degree of the gray values to be allocated and the number of the pixel points of the gray values to be allocated, judging the number of the pixel points to be allocated to obtain the number of the pixel points to be allocated of each gray value to be allocated; gray scale adjustment is carried out on the pixel points of the gray values to be allocated according to the number of the pixel points to be allocated to obtain a contrast gray scale histogram of each gray value to be allocated; obtaining contrast difference of each gray value to be allocated according to the difference between the gray value corresponding pixel points in the contrast gray histogram, wherein the contrast difference is used for describing the difference between the image blocks of the industrial material gray image and the image blocks of the industrial material gray image without noise interference;
screening ideal gray values from a plurality of gray values to be allocated according to the contrast difference; performing defect monitoring on the industrial gray denoising image according to the ideal gray value;
the method for screening out a plurality of main gray scale intervals formed by gray scale values to be distributed from intervals formed by all gray scale values in the image block comprises the following specific steps:
the method comprises the steps of marking the number of preset gray values as T2, marking the gray value with the largest number of pixel points in a gray histogram of an image block as a main body center gray value for any image block, marking a section formed by continuous T2 gray values with the number of pixels being different from 0 before the main body center gray value and continuous T2 gray values with the number of pixels being different from 0 after the main body center gray value as a main body gray section of the image block, and marking each gray value in the main body gray section as a gray value to be allocated;
the gray value range judgment is carried out according to the noise degree of the gray value to be allocated and the number of the gray values to be allocated in the main gray interval, so as to obtain the pixel point allocation range of each gray value to be allocated, and the specific method comprises the following steps:
for any one gray value to be allocated in the main gray interval of any one image block, in the formula,a pixel point allocation range representing a gray value to be allocated; />Representing the noise level of the gray value to be allocated; />Representing all the ashes to be distributed in the main gray scale intervalThe number of degree values; />Representing an upward rounding.
2. The industrial material quality monitoring method based on artificial intelligence and the internet of things according to claim 1, wherein the counting of the number of the connected domains of each gray value in the image block to obtain the gray connected domain duty ratio of each gray value comprises the following specific steps:
for any image block, carrying out connected domain analysis on the image block to obtain a plurality of connected domains of the image block, and obtaining a gray level histogram of the image block; for any gray value in the gray histogram of the image block, each connected domain of the gray value is marked as one gray connected domain of the gray value, all the gray connected domains of all the gray values are obtained, the number of the gray connected domains of all the gray values is subjected to linear normalization, and each normalized number is marked as the gray connected domain duty ratio of the gray value.
3. The industrial material quality monitoring method based on artificial intelligence and the internet of things according to claim 1, wherein the noise-containing degree of each gray value is obtained according to the gray connected domain duty ratio, comprising the following specific steps:
for any image block, carrying out connected domain analysis on the image block to obtain a plurality of connected domains of the image block, and obtaining a gray level histogram of the image block; for any gray value in the gray histogram of the image block, marking each connected domain of the gray value as one gray connected domain of the gray value;
for any gray connected domain, acquiring gradient amplitude values of all pixel points in the gray connected domain by using a sobel operator, and marking the average value of the gradient amplitude values of all pixel points in the gray connected domain as the gray gradient amplitude value of the gray connected domain;
in the method, in the process of the invention,the noise level of the gray value is represented; />A gray connected domain duty ratio representing a gray value; />Representing the variance of the gray gradient amplitude values of all gray connected domains under the gray value; />An exponential function based on a natural constant is represented.
4. The industrial material quality monitoring method based on artificial intelligence and the internet of things according to claim 1, wherein the method for determining the number of the pixel points according to the noise level of the gray value to be allocated and the number of the pixel points of the gray value to be allocated to obtain the number of the pixel points to be allocated of each gray value to be allocated comprises the following specific steps:
for any one gray value to be allocated in the main gray interval of any one image block, in the formula,representing the number of pixel points to be allocated of the gray value to be allocated; />Representing the noise level of the gray value to be allocated; />Representing the gray value to be allocated with the largest pixel number; />Representing a gray value to be assigned; />Representing the number of all gray values to be allocated in the main gray interval; />Representing the number of pixels to be assigned with gray values; />The representation takes absolute value; />Representing an upward rounding.
5. The industrial material quality monitoring method based on artificial intelligence and the internet of things according to claim 1, wherein the method for obtaining a contrast gray level histogram of each gray level value to be allocated by performing gray level adjustment on the pixel points of the gray level value to be allocated according to the number of the pixel points to be allocated comprises the following specific steps:
for any gray value to be allocated in the main gray interval of any image block, the gray value to be allocated is precededAfter the gray value and the gray value to be allocated +.>The gray values are all marked as gray values to be adjusted of the gray values to be allocated; randomly acquiring the number of pixel points to be allocated in the gray value to be allocated, and marking the pixel points to be allocated as the gray value to be allocated;
performing a gray scale adjustment operation, the gray scale adjustment operation comprising: the gray value of any pixel point to be allocated with the gray value which is not adjusted is adjusted to be the first gray value to be adjusted by taking the first gray value to be adjusted of the gray value to be allocated as a starting point; the gray value of any pixel point to be allocated which is not adjusted with the gray value is adjusted to be a second gray value to be adjusted; adjusting the gray value of any pixel point to be allocated with the gray value which is not adjusted to be a third gray value to be adjusted, and the like until the pixel point to be allocated with the gray value which is not adjusted is not existed or all the gray values to be adjusted are traversed;
if all the gray values to be adjusted are traversed, pixel points to be allocated with unadjusted gray values exist, and the gray adjustment operation is repeated until no pixel points to be allocated with unadjusted gray values exist;
and marking the corresponding gray level histogram after gray level adjustment of all the allocated pixel points of the gray level value to be allocated as a contrast gray level histogram of the gray level value to be allocated.
6. The industrial material quality monitoring method based on artificial intelligence and the internet of things according to claim 1, wherein the obtaining the contrast difference of each gray value to be allocated according to the difference between the gray value corresponding to the pixel point in the contrast gray histogram comprises the following specific steps:
for any one gray value to be allocated in the main gray interval of any one image block, in the formula,representing a contrast difference in gray values to be assigned; />The deviation of the number of the pixel points of all the gray values to be allocated in the main gray interval in the contrast gray histogram of the gray values to be allocated; />And the overstep kurtosis of the number of the pixel points of all the gray values to be allocated in the main gray interval in the contrast gray histogram of the gray values to be allocated.
7. The industrial material quality monitoring method based on artificial intelligence and the internet of things according to claim 1, wherein the method for screening out ideal gray values from a plurality of gray values to be distributed according to contrast difference comprises the following specific steps:
and for any image block, marking the gray value to be allocated with the largest contrast difference as an ideal gray value of the image block.
8. The industrial material quality monitoring method based on artificial intelligence and the internet of things according to claim 1, wherein the defect monitoring of the industrial gray-scale denoising image according to the ideal gray-scale value comprises the following specific steps:
for any one image block in any one industrial material gray level image, in a main gray level interval on a gray level histogram of the image block, marking the serial number of each gray level value to be allocated of the image block as the iteration times of the image block; the serial number of the ideal gray value of the image block is marked as the ideal iteration number of the image block;
in the method, in the process of the invention,filter regulation parameters representing image blocks: />Representing an ideal iteration number of the image block; />Representing the maximum value of all iteration times of the image block;
replacing smoothing parameters of wiener filtering with filtering regulation parameters of the image blocks, carrying out wiener filtering on the image blocks to obtain denoised image blocks, and marking the industrial material gray level image as an industrial material gray level denoising image after all denoised image blocks in the industrial material gray level image are obtained;
inputting the industrial material gray level denoising image into a trained neural network to obtain a plurality of defect areas in the industrial material gray level denoising image, and obtaining a plurality of defect areas of all the industrial gray level denoising images.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311525547.9A CN117237245B (en) | 2023-11-16 | 2023-11-16 | Industrial material quality monitoring method based on artificial intelligence and Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311525547.9A CN117237245B (en) | 2023-11-16 | 2023-11-16 | Industrial material quality monitoring method based on artificial intelligence and Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117237245A CN117237245A (en) | 2023-12-15 |
| CN117237245B true CN117237245B (en) | 2024-01-26 |
Family
ID=89097065
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311525547.9A Active CN117237245B (en) | 2023-11-16 | 2023-11-16 | Industrial material quality monitoring method based on artificial intelligence and Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117237245B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107730548A (en) * | 2017-10-23 | 2018-02-23 | 厦门诺银科技有限公司 | It is a kind of based on average gray and area towards heating furnace flame real-time detection method |
| WO2022141178A1 (en) * | 2020-12-30 | 2022-07-07 | 深圳市大疆创新科技有限公司 | Image processing method and apparatus |
| CN115100191A (en) * | 2022-08-22 | 2022-09-23 | 南通恒强轧辊有限公司 | Metal casting defect identification method based on industrial detection |
| CN116309757A (en) * | 2023-05-24 | 2023-06-23 | 山东省青东智能科技有限公司 | Binocular stereo matching method based on machine vision |
| CN116363133A (en) * | 2023-06-01 | 2023-06-30 | 无锡斯达新能源科技股份有限公司 | Illuminator accessory defect detection method based on machine vision |
| CN116645367A (en) * | 2023-07-27 | 2023-08-25 | 山东昌啸商贸有限公司 | Steel plate cutting quality detection method for high-end manufacturing |
| CN116721108A (en) * | 2023-08-11 | 2023-09-08 | 山东奥晶生物科技有限公司 | Stevioside product impurity detection method based on machine vision |
| CN116740065A (en) * | 2023-08-14 | 2023-09-12 | 山东伟国板业科技有限公司 | Quick tracing method and system for defective products of artificial board based on big data |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9430701B2 (en) * | 2014-02-07 | 2016-08-30 | Tata Consultancy Services Limited | Object detection system and method |
-
2023
- 2023-11-16 CN CN202311525547.9A patent/CN117237245B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107730548A (en) * | 2017-10-23 | 2018-02-23 | 厦门诺银科技有限公司 | It is a kind of based on average gray and area towards heating furnace flame real-time detection method |
| WO2022141178A1 (en) * | 2020-12-30 | 2022-07-07 | 深圳市大疆创新科技有限公司 | Image processing method and apparatus |
| CN115100191A (en) * | 2022-08-22 | 2022-09-23 | 南通恒强轧辊有限公司 | Metal casting defect identification method based on industrial detection |
| CN116309757A (en) * | 2023-05-24 | 2023-06-23 | 山东省青东智能科技有限公司 | Binocular stereo matching method based on machine vision |
| CN116363133A (en) * | 2023-06-01 | 2023-06-30 | 无锡斯达新能源科技股份有限公司 | Illuminator accessory defect detection method based on machine vision |
| CN116645367A (en) * | 2023-07-27 | 2023-08-25 | 山东昌啸商贸有限公司 | Steel plate cutting quality detection method for high-end manufacturing |
| CN116721108A (en) * | 2023-08-11 | 2023-09-08 | 山东奥晶生物科技有限公司 | Stevioside product impurity detection method based on machine vision |
| CN116740065A (en) * | 2023-08-14 | 2023-09-12 | 山东伟国板业科技有限公司 | Quick tracing method and system for defective products of artificial board based on big data |
Non-Patent Citations (2)
| Title |
|---|
| New image denoising algorithm based on improved grey prediction model;Songyun Xie 等;IEEE;全文 * |
| 基于多特征融合的有监督视网膜血管提取;梁礼明 等;计算机学报(第11期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117237245A (en) | 2023-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103295191A (en) | Multi-scale vision self-adaptation image enhancing method and evaluating method | |
| CN115984148B (en) | Denoising enhancement method for high-throughput gene sequencing data | |
| CN111192251B (en) | Follicle ultrasonic processing method and system based on level set image segmentation | |
| CN110288618B (en) | Multi-target segmentation method for uneven-illumination image | |
| CN117314801B (en) | Fuzzy image optimization enhancement method based on artificial intelligence | |
| CN113592782B (en) | Method and system for extracting X-ray image defects of composite material carbon fiber core rod | |
| CN116934763B (en) | Medical rubber plug defect detection method based on visual characteristics | |
| CN116703784B (en) | Heart ultrasonic image vision enhancement method | |
| CN116645363B (en) | Vision-based starch production quality real-time detection method | |
| CN114529538A (en) | Textile surface defect detection method based on artificial intelligence and Gaussian mixture model | |
| CN108765448B (en) | Shrimp larvae counting analysis method based on improved TV-L1 model | |
| CN116246174B (en) | Sweet potato variety identification method based on image processing | |
| CN117422712A (en) | Plastic master batch visual detection method and system based on image filtering processing | |
| CN115272303A (en) | Textile fabric defect degree evaluation method, device and system based on Gaussian blur | |
| CN117593193B (en) | Sheet metal image enhancement method and system based on machine learning | |
| CN116993764B (en) | Stomach CT intelligent segmentation extraction method | |
| CN117809379A (en) | Intelligent humanoid recognition alarm system and method based on monitoring camera | |
| CN117237245B (en) | Industrial material quality monitoring method based on artificial intelligence and Internet of things | |
| CN108133467B (en) | Underwater image enhancement system and method based on particle calculation | |
| CN113470058A (en) | Gravel particle size distribution measuring method and device | |
| CN117974528B (en) | Kidney biopsy slice image optimization enhancement method | |
| CN117173192B (en) | Intelligent detection method and system for pork quality | |
| CN112101377B (en) | Online intermittent hollow filter stick detection method based on regional feature analysis | |
| CN118096579B (en) | 3D printing lattice structure defect detection method | |
| CN116958134B (en) | Plastic film extrusion quality evaluation method based on image processing |
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 |