CN117474912A - Road section gap analysis method and model based on computer vision - Google Patents
Road section gap analysis method and model based on computer vision Download PDFInfo
- Publication number
- CN117474912A CN117474912A CN202311811131.3A CN202311811131A CN117474912A CN 117474912 A CN117474912 A CN 117474912A CN 202311811131 A CN202311811131 A CN 202311811131A CN 117474912 A CN117474912 A CN 117474912A
- Authority
- CN
- China
- Prior art keywords
- image
- gap
- deep learning
- learning model
- unet
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000012357 Gap analysis Methods 0.000 title claims abstract description 27
- 238000013136 deep learning model Methods 0.000 claims abstract description 65
- 238000012549 training Methods 0.000 claims abstract description 41
- 238000012545 processing Methods 0.000 claims abstract description 27
- 238000001514 detection method Methods 0.000 claims abstract description 25
- 238000000605 extraction Methods 0.000 claims abstract description 24
- 238000005516 engineering process Methods 0.000 claims abstract description 7
- 238000012360 testing method Methods 0.000 claims description 36
- 238000007781 pre-processing Methods 0.000 claims description 18
- 238000001914 filtration Methods 0.000 claims description 12
- 238000012797 qualification Methods 0.000 claims description 12
- 238000003708 edge detection Methods 0.000 claims description 10
- 230000000877 morphologic effect Effects 0.000 claims description 8
- 238000012800 visualization Methods 0.000 claims description 7
- 238000003672 processing method Methods 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims 1
- 238000004458 analytical method Methods 0.000 description 9
- 238000012423 maintenance Methods 0.000 description 6
- 238000005260 corrosion Methods 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000008961 swelling Effects 0.000 description 2
- 238000011282 treatment Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/0985—Hyperparameter optimisation; Meta-learning; Learning-to-learn
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30132—Masonry; Concrete
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/06—Recognition of objects for industrial automation
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a road section gap analysis method and model based on computer vision, relates to the technical field of image processing, aims at solving the problems that the traditional road section gap analysis method relies on manual detection and has low efficiency and is easy to generate errors, and adopts the following technical scheme: image acquisition and pretreatment are carried out on the appointed road section; extracting a pavement area from the preprocessed image by utilizing a computer vision technology, detecting gaps from the extracted pavement area, and extracting features of the detected gaps; marking the preprocessed image based on the gap feature extraction result, and training a deep learning model Unet++ by using the marked image; collecting a new road surface image, inputting a trained deep learning model Unet++, and outputting a gap classification result by the deep learning model Unet++; and visually displaying the gap detection result, the feature extraction result and the gap classification result. The invention can detect and analyze the gap of the appointed road section, and improve the working efficiency and the accuracy.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a road section gap analysis method and model based on computer vision.
Background
In road maintenance and servicing, gap analysis is an important task. The pavement gap can reflect the service condition and service life of the pavement, and has important influence on maintenance and maintenance of the pavement. The traditional pavement gap analysis method mainly relies on manual detection, so that the efficiency is low, errors are easy to occur, and the requirement of modern road maintenance cannot be met. In recent years, the rapid development of computer vision technology has provided new solutions for automated analysis of road gaps. However, the accuracy and robustness of existing computer vision methods in road gap detection and classification remain to be improved. Therefore, the invention provides a road section gap analysis method and model based on computer vision so as to improve the accuracy and the robustness of gap detection and classification.
Disclosure of Invention
Aiming at the needs and the shortcomings of the prior art development, the invention provides a road section gap analysis method and model based on computer vision, which are used for improving the accuracy and the robustness of gap detection and classification.
In a first aspect, the present invention provides a road section gap analysis method based on computer vision, which solves the above technical problems by adopting the following technical scheme:
a road section gap analysis method based on computer vision comprises the following steps:
s1, image acquisition and preprocessing are carried out on a specified road section;
s2, extracting a pavement area from the preprocessed image by utilizing a computer vision technology, detecting gaps from the extracted pavement area, and extracting features of the detected gaps;
s3, marking the preprocessed image based on the gap feature extraction result, and training a deep learning model Unet++ by using the marked image;
s4, acquiring a new road surface image, inputting a deep learning model Unet++ trained in the step S3, and outputting a gap classification result by the deep learning model Unet++;
and S5, visually displaying the gap detection result, the feature extraction result and the gap classification result of the step S4.
Optionally, the step S1 is executed, and the preprocessing operation performed on the image sequentially includes:
image denoising is carried out by adopting a median filtering or Gaussian filtering denoising method;
image enhancement is carried out by adopting a histogram equalization and contrast stretching image enhancement method;
and performing binarization processing on the image by adopting a binarization method of global threshold binarization and local self-adaptive binarization.
Optionally, step S2 is executed, extracting a pavement area from the preprocessed image by using an image processing method, detecting a gap from the extracted pavement area by sequentially using an edge detection algorithm, a morphological processing and an area growth algorithm, and extracting features of the gap withdrawn from the detection, where the extracted gap features include: position information, length information, width information, and direction information of the slit.
Optionally, the specific operation of step S3 includes:
firstly, manually marking the preprocessed image based on a feature extraction result;
then, dividing the marked image into a training set and a testing set, wherein the marking result of the training set comprises all categories of gaps;
then training the deep learning model Unet++ by using the training set, testing the trained deep learning model Unet++ by using unlabeled images corresponding to the testing set, comparing the labeled result of the testing set with the output result of the deep learning model Unet++,
if the test qualification rate of the deep learning model Unet++ exceeds the set threshold, outputting the deep learning model Unet++ for the step S4,
if the test qualification rate of the deep learning model Unet++ does not exceed the set threshold value, the sample number of the training set is enlarged, and the deep learning model Unet++ is retrained.
In a second aspect, the invention provides a road section gap analysis model based on computer vision, which solves the technical problems as follows:
a computer vision based road segment gap analysis model, comprising:
the image acquisition module is used for acquiring images of the appointed road section;
the image preprocessing module is used for preprocessing the acquired image;
the image processing module is used for extracting a pavement area from the preprocessed image by utilizing a computer vision technology, detecting gaps from the extracted pavement area and extracting features of the detected gaps; according to the feature extraction result, carrying out manual marking on the preprocessed image;
the model training module is used for training a deep learning model Unet++ by using the manually marked images;
the input/output module is used for inputting the newly acquired road surface image into a trained deep learning model Unet++, and sending a gap classification result output by the deep learning model Unet++ to the visualization module;
and the visualization module is used for carrying out visual display on the gap detection result, the feature extraction result and the gap classification result of the input/output module of the image processing module.
Optionally, the preprocessing operation performed on the acquired image by the related image preprocessing module sequentially includes:
image denoising is carried out by adopting a median filtering or Gaussian filtering denoising method;
image enhancement is carried out by adopting a histogram equalization and contrast stretching image enhancement method;
and performing binarization processing on the image by adopting a binarization method of global threshold binarization and local self-adaptive binarization.
Optionally, the image processing module extracts a pavement area from the preprocessed image by using an image processing method, detects a gap from the extracted pavement area by sequentially using an edge detection algorithm, morphological processing and an area growth algorithm, and performs feature extraction on the gap withdrawn from the detection, where the extracted gap feature includes: position information, length information, width information, and direction information of the slit.
Optionally, the related road section gap analysis model further comprises a data dividing module;
after the preprocessed image is marked manually according to the feature extraction result of the image processing module, the marked image is divided into a training set and a testing set by using the data dividing module, and the marking result of the training set comprises all the categories of gaps;
the related model training module uses a training set to train a deep learning model Unet++, uses unlabeled images corresponding to a test set to test the trained deep learning model Unet++, compares the labeled result of the test set with the output result of the deep learning model Unet++,
if the test qualification rate of the deep learning model Unet++ exceeds the set threshold, outputting the deep learning model Unet++ for the step S4,
if the test qualification rate of the deep learning model Unet++ does not exceed the set threshold value, the sample number of the training set is enlarged, and the deep learning model Unet++ is retrained.
The road section gap analysis method and model based on computer vision have the beneficial effects that compared with the prior art:
(1) The invention can automatically detect and analyze the gap of the appointed road section, improves the working efficiency and the accuracy, and provides an important reference basis for the maintenance and the maintenance of the road surface;
(2) The invention uses the deep learning model Unet++ to carry out high-precision gap detection on the road surface image, and has higher accuracy and recall rate; gaps with different dimensions can be sensed at the same time, so that gaps with different sizes and shapes can be better adapted; the global context information can be perceived, so that the correlation between the background information of the road surface image and the gaps can be better understood, and the accuracy of gap detection and classification can be improved; the automatic analysis of the pavement can be realized, the working efficiency and the accuracy are greatly improved, and the interference of human factors is reduced;
(3) The deep learning model Unet++ used in the invention can be used for end-to-end training, which means that the whole model can be trained under a unified frame, and the matching problem among a plurality of steps is avoided.
Drawings
FIG. 1 is a flow chart of a method according to a first embodiment of the invention;
fig. 2 is a block diagram of a module connection according to a second embodiment of the present invention.
Detailed Description
In order to make the technical scheme, the technical problems to be solved and the technical effects of the invention more clear, the technical scheme of the invention is clearly and completely described below by combining specific embodiments.
Embodiment one:
referring to fig. 1, the present embodiment provides a road section gap analysis method based on computer vision, which includes the following steps:
s1, image acquisition and preprocessing are carried out on the appointed road section.
And (3) performing image acquisition on the appointed road section by using equipment such as a high-resolution camera or an unmanned aerial vehicle and the like to acquire a high-definition image of the road surface.
The preprocessing operation for the acquired image is as follows:
and denoising the image. The captured image may be noisy due to the effects of the photographing device or environmental factors. The image denoising can remove noise and improve the definition and quality of the image. Common denoising methods include median filtering, gaussian filtering, and the like.
And (5) enhancing the image. Due to the limitations of actual road conditions and shooting equipment, the acquired images may have the problems of uneven illumination, low contrast and the like. The image enhancement can improve the definition and contrast of the image by adjusting parameters such as brightness, contrast and the like, so that the gap is more obvious. Common image enhancement methods include histogram equalization, contrast stretching, and the like.
And (5) binarizing the image. In order to facilitate subsequent gap detection and classification, the image needs to be binarized. Binarization can convert an image into a black-and-white binary image, so that gaps and a background are more clear. Common binarization methods include global threshold binarization, local adaptive binarization, and the like.
S2, extracting a pavement area from the preprocessed image by using an image processing method, detecting gaps from the extracted pavement area by using an edge detection algorithm, morphological processing and an area growth algorithm in sequence, and extracting features of the gaps withdrawn from detection.
Edge detection may detect gaps by identifying edges in the image. Common edge detection algorithms include Sobel, canny, and the like. By detecting edges in the image, a gap on the road surface can be initially identified.
Morphological treatments may further enhance edges and remove noise by operations such as swelling, corrosion, etc. The operations can enhance the contrast of the gap, remove some irrelevant details and improve the accuracy of gap detection.
The region growing algorithm is a pixel-based segmentation method that can segment an image into different regions according to the similarity between pixels. In gap detection, a seed pixel may be defined and then adjacent pixels identified by region growing, thereby forming a complete gap.
The extracted slit features include: position information, length information, width information, and direction information of the slit. The position information of the gap, i.e. the position information of the gap in the road surface area, includes an abscissa and an ordinate. From the position information, the relative and absolute positions of the slit on the road surface can be determined. The length information of the gap, namely the pixel length from the starting point to the end point of the gap, can be used for judging the length of the gap and provides basis for subsequent classification and analysis. The width information of the gap, namely the width of the pixels in the vertical direction of the gap, can be used for judging the width of the gap, and provides basis for subsequent classification and analysis. The direction information of the gap, namely the direction angle of the gap, can be used for judging the direction of the gap and provides basis for subsequent classification and analysis.
S3, marking the preprocessed image based on the gap feature extraction result, and training a deep learning model Unet++ by using the marked image, wherein the specific operation comprises the following steps:
firstly, manually marking the preprocessed image based on a feature extraction result;
then, dividing the marked image into a training set and a testing set, wherein the marking result of the training set comprises all categories of gaps;
then training the deep learning model Unet++ by using the training set, testing the trained deep learning model Unet++ by using unlabeled images corresponding to the testing set, comparing the labeled result of the testing set with the output result of the deep learning model Unet++,
if the test qualification rate of the deep learning model Unet++ exceeds the set threshold, outputting the deep learning model Unet++ for the step S4,
if the test qualification rate of the deep learning model Unet++ does not exceed the set threshold value, the sample number of the training set is enlarged, and the deep learning model Unet++ is retrained.
S4, inputting the slit characteristics extracted in the step S2 into the deep learning model Unet++ trained in the step S3, and outputting a slit classification result by the deep learning model Unet++.
And S5, visually displaying the gap detection result, the feature extraction result and the gap classification result of the step S4.
The visualization may include various forms, such as labeling the gaps on the road surface, graphically displaying the classification results of the gaps.
Embodiment two:
referring to fig. 2, this embodiment proposes a road section gap analysis model based on computer vision, which includes:
the image acquisition module is used for acquiring images of the appointed road section;
the image preprocessing module is used for preprocessing the acquired image;
the image processing module is used for extracting a pavement area from the preprocessed image by utilizing a computer vision technology, detecting gaps from the extracted pavement area and extracting features of the detected gaps; according to the feature extraction result, carrying out manual marking on the preprocessed image;
the model training module is used for training a deep learning model Unet++ by using the manually marked images;
the input/output module is used for inputting the newly acquired road surface image into a trained deep learning model Unet++, and sending a gap classification result output by the deep learning model Unet++ to the visualization module;
and the visualization module is used for carrying out visual display on the gap detection result, the feature extraction result and the gap classification result of the input/output module of the image processing module.
In this embodiment, the image acquisition module performs image acquisition on the specified road section by using a high-resolution camera or an unmanned aerial vehicle and other devices, and obtains a high-definition image of the road surface.
The related preprocessing operation of the image preprocessing module on the acquired image sequentially comprises the following steps:
image denoising is carried out by adopting a median filtering or Gaussian filtering denoising method;
image enhancement is carried out by adopting a histogram equalization and contrast stretching image enhancement method;
and performing binarization processing on the image by adopting a binarization method of global threshold binarization and local self-adaptive binarization.
In this embodiment, the image processing module extracts the road surface area from the preprocessed image by using the image processing method, detects the gap from the extracted road surface area by sequentially using the edge detection algorithm, the morphological processing and the area growth algorithm, and performs feature extraction on the gap withdrawn from the detection.
Edge detection may detect gaps by identifying edges in the image. Common edge detection algorithms include Sobel, canny, and the like. By detecting edges in the image, a gap on the road surface can be initially identified.
Morphological treatments may further enhance edges and remove noise by operations such as swelling, corrosion, etc. The operations can enhance the contrast of the gap, remove some irrelevant details and improve the accuracy of gap detection.
The region growing algorithm is a pixel-based segmentation method that can segment an image into different regions according to the similarity between pixels. In gap detection, a seed pixel may be defined and then adjacent pixels identified by region growing, thereby forming a complete gap.
The gap characteristics extracted by the related image processing module comprise: position information, length information, width information, and direction information of the slit. The position information of the gap, i.e. the position information of the gap in the road surface area, includes an abscissa and an ordinate. From the position information, the relative and absolute positions of the slit on the road surface can be determined. The length information of the gap, namely the pixel length from the starting point to the end point of the gap, can be used for judging the length of the gap and provides basis for subsequent classification and analysis. The width information of the gap, namely the width of the pixels in the vertical direction of the gap, can be used for judging the width of the gap, and provides basis for subsequent classification and analysis. The direction information of the gap, namely the direction angle of the gap, can be used for judging the direction of the gap and provides basis for subsequent classification and analysis.
In this embodiment, the related road section gap analysis model further includes a data dividing module;
after the preprocessed image is marked manually according to the feature extraction result of the image processing module, the marked image is divided into a training set and a testing set by using the data dividing module, and the marking result of the training set comprises all the categories of gaps;
then, the model training module trains the deep learning model Unet++ by using the training set, tests the trained deep learning model Unet++ by using the unlabeled preprocessed image corresponding to the test set, compares the labeling result of the test set with the output result of the deep learning model Unet++,
if the test qualification rate of the deep learning model Unet++ exceeds the set threshold, outputting the deep learning model Unet++ for the step S4,
if the test qualification rate of the deep learning model Unet++ does not exceed the set threshold value, the sample number of the training set is enlarged, and the deep learning model Unet++ is retrained.
In summary, by adopting the road section gap analysis method and model based on computer vision, gap detection and analysis can be automatically performed on the appointed road section, and the working efficiency and accuracy can be improved.
The foregoing has outlined rather broadly the principles and embodiments of the present invention in order that the detailed description of the invention may be better understood. Based on the above-mentioned embodiments of the present invention, any improvements and modifications made by those skilled in the art without departing from the principles of the present invention should fall within the scope of the present invention.
Claims (10)
1. The road section gap analysis method based on computer vision is characterized by comprising the following steps of:
s1, image acquisition and preprocessing are carried out on a specified road section;
s2, extracting a pavement area from the preprocessed image by utilizing a computer vision technology, detecting gaps from the extracted pavement area, and extracting features of the detected gaps;
s3, marking the preprocessed image based on the gap feature extraction result, and training a deep learning model Unet++ by using the marked image;
s4, acquiring a new road surface image, inputting a deep learning model Unet++ trained in the step S3, and outputting a gap classification result by the deep learning model Unet++;
and S5, visually displaying the gap detection result, the feature extraction result and the gap classification result of the step S4.
2. The method for analyzing road section gaps based on computer vision according to claim 1, wherein the step S1 is performed, and the preprocessing operation performed on the image is as follows:
image denoising is carried out by adopting a median filtering or Gaussian filtering denoising method;
image enhancement is carried out by adopting a histogram equalization and contrast stretching image enhancement method;
and performing binarization processing on the image by adopting a binarization method of global threshold binarization and local self-adaptive binarization.
3. The computer vision-based road segment gap analysis method according to claim 1, wherein step S2 is performed, wherein the road surface region is extracted from the preprocessed image by using an image processing method, and the gap is detected from the extracted road surface region by sequentially using an edge detection algorithm, a morphological processing, and a region growing algorithm, and the feature extraction is performed on the gap withdrawn from the detection.
4. A computer vision based road segment gap analysis method according to claim 3, characterized in that the extracted gap features comprise: position information, length information, width information, and direction information of the slit.
5. The method for analyzing road section gaps based on computer vision according to claim 1, wherein the specific operations of step S3 include:
firstly, manually marking the preprocessed image based on a feature extraction result;
then, dividing the marked image into a training set and a testing set, wherein the marking result of the training set comprises all categories of gaps;
then training the deep learning model Unet++ by using the training set, testing the trained deep learning model Unet++ by using unlabeled images corresponding to the testing set, comparing the labeled result of the testing set with the output result of the deep learning model Unet++,
if the test qualification rate of the deep learning model Unet++ exceeds the set threshold, outputting the deep learning model Unet++ for the step S4,
if the test qualification rate of the deep learning model Unet++ does not exceed the set threshold value, the sample number of the training set is enlarged, and the deep learning model Unet++ is retrained.
6. A computer vision-based road segment gap analysis model, comprising:
the image acquisition module is used for acquiring images of the appointed road section;
the image preprocessing module is used for preprocessing the acquired image;
the image processing module is used for extracting a pavement area from the preprocessed image by utilizing a computer vision technology, detecting gaps from the extracted pavement area and extracting features of the detected gaps; according to the feature extraction result, carrying out manual marking on the preprocessed image;
the model training module is used for training a deep learning model Unet++ by using the manually marked images;
the input/output module is used for inputting the newly acquired road surface image into a trained deep learning model Unet++, and sending a gap classification result output by the deep learning model Unet++ to the visualization module;
and the visualization module is used for carrying out visual display on the gap detection result, the feature extraction result and the gap classification result of the input/output module of the image processing module.
7. The computer vision-based road segment gap analysis model of claim 6, wherein the preprocessing operation performed by the image preprocessing module on the acquired image sequentially comprises:
image denoising is carried out by adopting a median filtering or Gaussian filtering denoising method;
image enhancement is carried out by adopting a histogram equalization and contrast stretching image enhancement method;
and performing binarization processing on the image by adopting a binarization method of global threshold binarization and local self-adaptive binarization.
8. The computer vision-based road segment gap analysis model according to claim 6, wherein the image processing module extracts road surface regions from the preprocessed images by using an image processing method, detects gaps from the extracted road surface regions by using an edge detection algorithm, a morphological processing and a region growing algorithm in sequence, and performs feature extraction on the detected and withdrawn gaps.
9. The computer vision-based road segment gap analysis model of claim 8, wherein the gap features extracted by the image processing module comprise: position information, length information, width information, and direction information of the slit.
10. The computer vision based road segment gap analysis model of claim 6, further comprising a data partitioning module;
after the preprocessed image is marked manually according to the feature extraction result of the image processing module, the marked image is divided into a training set and a testing set by using the data dividing module, and the marking result of the training set comprises all the categories of gaps;
the model training module trains the deep learning model Unet++ by using a training set, tests the trained deep learning model Unet++ by using unlabeled images corresponding to a testing set, compares the labeled result of the testing set with the output result of the deep learning model Unet++,
if the test qualification rate of the deep learning model Unet++ exceeds the set threshold, outputting the deep learning model Unet++ for the step S4,
if the test qualification rate of the deep learning model Unet++ does not exceed the set threshold value, the sample number of the training set is enlarged, and the deep learning model Unet++ is retrained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311811131.3A CN117474912A (en) | 2023-12-27 | 2023-12-27 | Road section gap analysis method and model based on computer vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311811131.3A CN117474912A (en) | 2023-12-27 | 2023-12-27 | Road section gap analysis method and model based on computer vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117474912A true CN117474912A (en) | 2024-01-30 |
Family
ID=89635122
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311811131.3A Pending CN117474912A (en) | 2023-12-27 | 2023-12-27 | Road section gap analysis method and model based on computer vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117474912A (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107862677A (en) * | 2017-10-16 | 2018-03-30 | 中铁第四勘察设计院集团有限公司 | The Tunnel Lining Cracks recognition methods of thresholding algorithm and system between a kind of class based on gradient |
| CN109632103A (en) * | 2018-11-22 | 2019-04-16 | 西安理工大学 | High vacant building Temperature Distribution and surface crack remote supervision system and monitoring method |
| CN112508030A (en) * | 2020-12-18 | 2021-03-16 | 山西省信息产业技术研究院有限公司 | Tunnel crack detection and measurement method based on double-depth learning model |
| CN112907561A (en) * | 2021-03-16 | 2021-06-04 | 中科海拓(无锡)科技有限公司 | Notebook appearance flaw detection method based on deep learning |
| CN114359542A (en) * | 2021-11-26 | 2022-04-15 | 广州大学 | Concrete structure crack identification method based on computer vision and deep learning |
| CN114511014A (en) * | 2022-01-21 | 2022-05-17 | 北京城建勘测设计研究院有限责任公司 | Subway tunnel water leakage detection system and method based on image deep learning algorithm |
| CN114596551A (en) * | 2022-03-03 | 2022-06-07 | 湘潭大学 | Vehicle-mounted forward-looking image crack detection method |
| CN114998852A (en) * | 2021-08-05 | 2022-09-02 | 浙江杉工智能科技有限公司 | Intelligent detection method for road pavement diseases based on deep learning |
| CN116228757A (en) * | 2023-05-08 | 2023-06-06 | 山东省海洋科学研究院(青岛国家海洋科学研究中心) | Deep sea cage and netting detection method based on image processing algorithm |
| CN116385430A (en) * | 2023-05-26 | 2023-07-04 | 浙江视觉智能创新中心有限公司 | Machine vision flaw detection method, device, medium and equipment |
| CN116434206A (en) * | 2023-03-20 | 2023-07-14 | 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) | Cotton quality character detection method based on machine vision technology |
| CN117237736A (en) * | 2023-10-09 | 2023-12-15 | 四川轻化工大学 | Daqu quality detection method based on machine vision and deep learning |
-
2023
- 2023-12-27 CN CN202311811131.3A patent/CN117474912A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107862677A (en) * | 2017-10-16 | 2018-03-30 | 中铁第四勘察设计院集团有限公司 | The Tunnel Lining Cracks recognition methods of thresholding algorithm and system between a kind of class based on gradient |
| CN109632103A (en) * | 2018-11-22 | 2019-04-16 | 西安理工大学 | High vacant building Temperature Distribution and surface crack remote supervision system and monitoring method |
| CN112508030A (en) * | 2020-12-18 | 2021-03-16 | 山西省信息产业技术研究院有限公司 | Tunnel crack detection and measurement method based on double-depth learning model |
| CN112907561A (en) * | 2021-03-16 | 2021-06-04 | 中科海拓(无锡)科技有限公司 | Notebook appearance flaw detection method based on deep learning |
| CN114998852A (en) * | 2021-08-05 | 2022-09-02 | 浙江杉工智能科技有限公司 | Intelligent detection method for road pavement diseases based on deep learning |
| CN114359542A (en) * | 2021-11-26 | 2022-04-15 | 广州大学 | Concrete structure crack identification method based on computer vision and deep learning |
| CN114511014A (en) * | 2022-01-21 | 2022-05-17 | 北京城建勘测设计研究院有限责任公司 | Subway tunnel water leakage detection system and method based on image deep learning algorithm |
| CN114596551A (en) * | 2022-03-03 | 2022-06-07 | 湘潭大学 | Vehicle-mounted forward-looking image crack detection method |
| CN116434206A (en) * | 2023-03-20 | 2023-07-14 | 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) | Cotton quality character detection method based on machine vision technology |
| CN116228757A (en) * | 2023-05-08 | 2023-06-06 | 山东省海洋科学研究院(青岛国家海洋科学研究中心) | Deep sea cage and netting detection method based on image processing algorithm |
| CN116385430A (en) * | 2023-05-26 | 2023-07-04 | 浙江视觉智能创新中心有限公司 | Machine vision flaw detection method, device, medium and equipment |
| CN117237736A (en) * | 2023-10-09 | 2023-12-15 | 四川轻化工大学 | Daqu quality detection method based on machine vision and deep learning |
Non-Patent Citations (2)
| Title |
|---|
| 王保宪 等: "基于特征增强学习的路面裂缝病害视觉检测方法", 铁道科学与工程学报, vol. 19, no. 7, 31 July 2022 (2022-07-31), pages 1927 - 1935 * |
| 边琛: "基于典型裂纹的沥青路面损坏状况视觉评估研究", 中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑, vol. 2022, no. 5, 15 May 2022 (2022-05-15), pages 034 - 92 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114937055B (en) | Image self-adaptive segmentation method and system based on artificial intelligence | |
| CN107257161B (en) | Transformer substation disconnecting link remote control auxiliary checking method and system based on state recognition algorithm | |
| CN113592861B (en) | Bridge crack detection method based on dynamic threshold | |
| CN109489724B (en) | Tunnel train safe operation environment comprehensive detection device and detection method | |
| CN111260616A (en) | Insulator crack detection method based on Canny operator two-dimensional threshold segmentation optimization | |
| CN111814686A (en) | Vision-based power transmission line identification and foreign matter invasion online detection method | |
| Yimyam et al. | The automated parking fee calculation using license plate recognition system | |
| CN108982512A (en) | A kind of circuit board detecting system and method based on machine vision | |
| CN109472788B (en) | Method for detecting flaw on surface of airplane rivet | |
| CN105044122A (en) | Copper part surface defect visual inspection system and inspection method based on semi-supervised learning model | |
| CN103729856B (en) | A kind of Fabric Defects Inspection detection method utilizing S-transformation signal extraction | |
| CN110648330B (en) | Defect detection method for camera glass | |
| CN109359554A (en) | A kind of forest fire recognition methods based on more composograph processing techniques | |
| CN112233111A (en) | Tunnel gap detection method based on digital image processing | |
| CN105718964A (en) | Transmission line vibration damper visual detection method | |
| CN107330440B (en) | Ocean state calculation method based on image recognition | |
| CN117330582A (en) | Polymer PE film surface crystal point detecting system | |
| CN108009480A (en) | A kind of image human body behavioral value method of feature based identification | |
| CN102313740A (en) | Solar panel crack detection method | |
| CN112330633B (en) | Jumper wire adhesive tape damage fault image segmentation method based on self-adaptive band-pass filtering | |
| CN117474912A (en) | Road section gap analysis method and model based on computer vision | |
| Huang et al. | Mango surface defect detection based on HALCON | |
| CN114494142A (en) | Mobile terminal middle frame defect detection method and device based on deep learning | |
| CN113642473A (en) | Mining coal machine state identification method based on computer vision | |
| CN113947563A (en) | Cable process quality dynamic defect detection method based on deep learning |
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 |