CN112801972A - Bridge defect detection method, device, system and storage medium - Google Patents
Bridge defect detection method, device, system and storage medium Download PDFInfo
- Publication number
- CN112801972A CN112801972A CN202110096158.3A CN202110096158A CN112801972A CN 112801972 A CN112801972 A CN 112801972A CN 202110096158 A CN202110096158 A CN 202110096158A CN 112801972 A CN112801972 A CN 112801972A
- Authority
- CN
- China
- Prior art keywords
- bridge
- image
- image data
- defect detection
- defect
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 109
- 238000001514 detection method Methods 0.000 title claims abstract description 65
- 238000012545 processing Methods 0.000 claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 33
- 238000003708 edge detection Methods 0.000 claims abstract description 13
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 230000011218 segmentation Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 claims description 7
- 238000009432 framing Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 5
- 239000013589 supplement Substances 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000033001 locomotion Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000005315 distribution function Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 235000002566 Capsicum Nutrition 0.000 description 1
- 239000006002 Pepper Substances 0.000 description 1
- 235000016761 Piper aduncum Nutrition 0.000 description 1
- 235000017804 Piper guineense Nutrition 0.000 description 1
- 244000203593 Piper nigrum Species 0.000 description 1
- 235000008184 Piper nigrum Nutrition 0.000 description 1
- 208000037656 Respiratory Sounds Diseases 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000036544 posture Effects 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- 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
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
-
- 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
- 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
- G06T2207/30181—Earth observation
- G06T2207/30184—Infrastructure
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a bridge defect detection method, equipment, a system and a storage medium, wherein the method comprises the following steps: acquiring bridge image data acquired by an unmanned aerial vehicle; preprocessing the bridge image data to obtain a clear image with a background except for the defect removed; performing edge detection on the image obtained after pretreatment to detect the area of the bridge body with peeling defects; and processing the preprocessed image by adopting a preset convolutional neural network model so as to detect the position of the crack defect of the bridge body and the size of the crack defect. The invention solves the problems of low working efficiency, long operation period and high risk in the conventional bridge defect detection.
Description
Technical Field
The invention relates to the technical field of bridge detection, in particular to a bridge defect detection method, equipment, a system and a storage medium.
Background
The bridge is used as an important component of traffic and is closely related to politics, economy, military, scientific technology, culture, art and the like. The safe operation of the bridge has close relation with national economy and people's life. The construction of bridges plays a crucial role in the development of national economy, but bridge decks are often affected by natural climate, and particularly in areas with heavy rainfall, the deck decks are more easily damaged by water stagnation. Meanwhile, with the economic development of China, the traffic flow and the vehicle load of the bridge are gradually increased, and the factors can increase the damage to the surface of the bridge. With the lapse of time, the bridge will produce problems such as crackle, road surface subside, local damage, shorten its life, even lead to serious accident such as bridge collapse. Therefore, frequent defect detection of the bridge is required.
And the traditional bridge detection and maintenance work has the defects of low working efficiency, long operation period, large risk and the like.
Disclosure of Invention
In view of the above, it is desirable to provide a bridge defect detecting method, device, system and storage medium, which are used to solve the problems of low working efficiency, long working cycle and high risk in the current bridge defect detection.
In a first aspect, the present invention provides a bridge defect detection method, including the following steps:
acquiring bridge image data acquired by an unmanned aerial vehicle;
preprocessing the bridge image data to obtain a clear image with a background except for the defect removed;
performing edge detection on the image obtained after pretreatment to detect the area of the bridge body with peeling defects;
and processing the preprocessed image by adopting a preset convolutional neural network model so as to detect the position of the crack defect of the bridge body and the size of the crack defect.
Preferably, in the method for detecting a bridge defect, the bridge body image data at least includes bridge deck image data, bridge frame image data, and pier image data.
Preferably, in the bridge defect detection method, the step of preprocessing the bridge image data to obtain a clear image with a background except for a defect removed includes:
performing framing processing on the bridge image data to obtain a plurality of key frame images;
and processing each key frame image by adopting an image processing technology to obtain a plurality of clear images with the background except the defects removed.
Preferably, in the bridge defect detection method, the step of performing framing processing on the bridge image data to obtain a plurality of key frame images specifically includes:
acquiring the total frame number of the bridge image data, reading the bridge image data frame by frame, and taking an image with the number of the frame number being N times of a preset value as a key frame image; wherein N is a natural number not less than 1.
Preferably, in the bridge defect detecting method, the step of processing each key frame image by using an image processing technique to obtain a plurality of clear images with the background except for the defect removed includes:
performing graying processing on the key frame image, and converting an image in an RGB domain into an image in a GRY domain;
equalizing the gray level of the image subjected to gray level processing through a histogram;
performing median filtering on the equalized image by adopting a median filtering method;
performing edge enhancement processing on the filtered image by adopting a Laplace algorithm;
and performing background segmentation on the image subjected to edge enhancement processing by adopting a threshold segmentation method to obtain a clear image with the background except the defect removed.
Preferably, in the bridge defect detecting method, the step of performing edge detection on the image obtained after the preprocessing to detect the region of the bridge body with the peeling defect specifically includes:
adopting a Canny edge detection algorithm to carry out edge extraction on the image, and searching a maximum connected domain of the edge after the edge is obtained;
and calculating the area of the maximum connected domain, comparing the area of the maximum connected domain with a preset area threshold, and judging that the maximum connected domain is a region needing to be repaired when the area of the maximum connected domain is larger than the preset area threshold.
Preferably, in the bridge defect detection method, the preset convolutional neural network model is a network model based on a YOLOV5 neural network framework.
In a second aspect, the invention further provides bridge defect detection equipment, a processor and a memory;
the memory has stored thereon a computer readable program executable by the processor;
the processor, when executing the computer readable program, implements the steps in the bridge defect detection method as described above.
In a third aspect, the present invention also provides a computer readable storage medium storing one or more programs, which are executable by one or more processors to implement the steps in the bridge defect detecting method as described above.
In a fourth aspect, the present invention further provides a bridge defect detecting system, including the above bridge defect detecting apparatus, further including:
the unmanned aerial vehicle is electrically connected with the bridge defect detection equipment and is used for acquiring bridge body image data;
the repairing vehicle is electrically connected with the bridge defect detection equipment and is used for overhauling according to the detection result of the bridge defect detection equipment;
and the light sources are respectively positioned on the unmanned aerial vehicle and the repairing vehicle and used for light supplement.
Compared with the prior art, the bridge defect detection method, the bridge defect detection equipment, the bridge defect detection system and the bridge defect storage medium provided by the invention have the advantages that the image is shot by the unmanned aerial vehicle, then the image is preprocessed, the edge detection method is used for detecting the peeling defect, the neural network model is used for detecting the crack defect, the detection precision is higher, the working efficiency is high, the working period is short, and the safety risk cannot be generated.
Drawings
FIG. 1 is a flowchart illustrating a bridge defect detection method according to a preferred embodiment of the present invention;
FIG. 2 is a diagram illustrating an effect of a pre-processed image according to a preferred embodiment of the present invention;
FIG. 3 is a diagram illustrating an effect of another preferred embodiment of the preprocessed image in the bridge defect detection method according to the present invention;
FIG. 4 is a diagram of a first result of convolutional neural network model identification in the bridge defect detection method provided by the present invention;
fig. 5 is a second result diagram of convolutional neural network model identification in the bridge defect detection method provided by the present invention.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
Referring to fig. 1, a bridge defect detection method according to an embodiment of the present invention includes the following steps:
s100, bridge image data acquired by the unmanned aerial vehicle are acquired.
In this embodiment, utilize the motion camera that carries on the unmanned aerial vehicle to come real-time acquisition pontic image data, pontic image data exists with the form of video stream, then through 5G technique with unmanned aerial vehicle communication, and then acquire the pontic image data that unmanned aerial vehicle gathered, wherein, pontic image data includes bridge floor image data, crane span structure image data and pier image data at least to realize the no dead angle defect detection to the pontic.
S200, preprocessing the bridge image data to obtain a clear image with the background except the defects removed.
In this embodiment, in order to ensure the accuracy of subsequent image recognition, the image data of the bridge body needs to be preprocessed to obtain a clear image, and the background other than the defect is filtered out, so that the recognition is convenient. In specific implementation, the step S200 specifically includes:
performing framing processing on the bridge image data to obtain a plurality of key frame images;
and processing each key frame image by adopting an image processing technology to obtain a plurality of clear images with the background except the defects removed.
Specifically, after the bridge image data is obtained through the 5G technology, the video stream is divided into different key frame images through framing processing, and then each key frame image is processed through the image processing technology, so that the images reach the expected definition, and the background except the defect is removed, and then the images are stored in the AVI format. The processing procedures of the image processing technology comprise processing procedures of graying processing, filtering processing, denoising processing, threshold segmentation processing and the like on the image so as to reduce the interference of the background on identification.
In a further embodiment, the step of performing framing processing on the bridge image data to obtain a plurality of key frame images specifically includes:
acquiring the total frame number of the bridge image data, reading the bridge image data frame by frame, and taking an image with the number of the frame number being N times of a preset value as a key frame image; wherein N is a natural number not less than 1.
Specifically, the total frame number of the image is first acquired by using the relevant OPENCV function, and this value is set as a threshold; reading the video of the video stream frame by frame, adding one to the count every reading one frame, dividing the frame number by 10 to obtain the processed image of every five frames, and storing the processed image to the designated storage address. The video stream thus gets the corresponding key frames according to the processing.
In a further embodiment, the step of processing each of the key frame images by using an image processing technique to obtain a plurality of clear images with the background other than the defect removed includes:
performing graying processing on the key frame image, and converting an image in an RGB domain into an image in a GRY domain;
equalizing the gray level of the image subjected to gray level processing through a histogram;
performing median filtering on the equalized image by adopting a median filtering method;
performing edge enhancement processing on the filtered image by adopting a Laplace algorithm;
and performing background segmentation on the image subjected to edge enhancement processing by adopting a threshold segmentation method to obtain a clear image with the background except the defect removed.
In this embodiment, since the key frame image is an RGB color image, the image needs to process three RGB components during processing, but the three-channel image only optically adjusts the color and does not play a key role in reflecting the essential features of the image. According to the invention, other unneeded backgrounds are inevitably acquired at the same time when the bridge is processed, so that a desired image which can better reflect the image characteristics is obtained by performing necessary steps such as threshold segmentation and the like conveniently.
The method is to reorder the pixels in a certain template from small to large, for each pixel, a window W composed of odd pixels is determined, the movement of the template window rearranges the pixels of the pixel block covered in the window, and the brightest or darkest pixels are arranged at both sides. The gray value of the middle position is assigned to the original pixel in the middle, so that a new pixel value is obtained, and similarly, the method can traverse all the pixels in the image once in the process of continuous movement, so that a new image is obtained. Therefore, salt and pepper noise can be removed easily, and original details and relevant edge features of the image are kept well.
Further, in order to prevent the gray scale conversion from affecting the position distribution of the original pixels and increase or decrease the number of pixels, the gray scale of the image after the gray scale processing needs to be equalized by using a histogram. The cumulative distribution function of each gray level is calculated by calculating the proportion, namely the probability percentage, of each gray level, and the corresponding gray level interval after change is judged according to the cumulative distribution function. The pixel levels in the transformed image are distributed evenly. The effect of improving the contrast of the image is achieved, and the specific effect is shown in fig. 2 and fig. 3, and it can be seen that the histogram equalization obviously enhances the valuable part of the image.
After equalization processing, 3 × 3 template blocks are set, gray values of all elements of each module are arranged from small to large through movement of the template blocks, and a median value is given to a central element of the template blocks, so that median filtering is achieved.
Furthermore, because a large number of noise points exist on an image, the noise points are distributed roughly and the edges of the noise points are blurred, the shadow contours of the bridge cracks can be interfered, and errors are brought to the subsequent length and width measurement, so that the edge contours need to be further processed, a laplacian-operator-based high-pass filter is constructed by selecting the laplacian operator with a good enhanced edge and introducing the laplacian operator, high-energy signals pass through the high-energy filter, and lower noise points are filtered.
After the edge enhancement processing is finished, the background segmentation is realized by adopting a threshold segmentation method, and when the threshold value of the image is set, the traditional method can cause the low-gray-level outline of the image part to be changed into an area with a gray value of 0, so that some bridge crack data can be lost, and the image information can be distorted. Therefore, the OTSU threshold is adopted, the threshold is automatically selected to be matched with image display, and Trackbar is set to adjust the target image in real time so as to ensure the influence of brightness change of different pictures on image threshold division.
S300, carrying out edge detection on the image obtained after preprocessing to detect the area of the bridge body with the peeling defect.
In this embodiment, the purpose of the edge detection is to determine a bridge peeling defect, and specifically, the step S300 specifically includes:
adopting a Canny edge detection algorithm to carry out edge extraction on the image, and searching a maximum connected domain of the edge after the edge is obtained;
and calculating the area of the maximum connected domain, comparing the area of the maximum connected domain with a preset area threshold, and judging that the maximum connected domain is a region needing to be repaired when the area of the maximum connected domain is larger than the preset area threshold.
In other words, the method utilizes a Canny edge detection algorithm to extract the contour of the bridge floor, obtains a closed connected domain through the obtained contour, compares the closed connected domain with a preset threshold, can judge that the bridge floor traffic is influenced and needs to be repaired if the closed connected domain is larger than the preset threshold, and marks the contour according to the result.
S400, processing the preprocessed image by adopting a preset convolutional neural network model to detect the position of the crack defect of the bridge body and the size of the crack defect.
Specifically, the preset convolutional neural network model is a network model based on a YOLOV5 neural network framework. The method utilizes a YOLOV5 neural network framework to judge and predict the preprocessed image, finds out cracks which meet the defect characteristics in the image and marks the cracks. Before the crack is automatically detected, the number of the convolution layers and the number of the full-connection layers need to be set according to the road surface condition, an optimal structure is designed, the minimum cost loss function is found through continuous iteration, the minimum cost loss function is enabled to be minimum, and the crack meeting the conditions is further obtained. Wherein the defect characteristics include length, width, shape characteristics, gray level threshold, etc. of the defect.
Further, when a convolutional neural network model is built, a data set is prepared, the data set utilizes pictures which are shot by a surveying unmanned aerial vehicle and serve as continuous image streams, the data set is divided into a training set and a verification set, about 5000 images which are framed in different postures and different bridge cracks are collected, 4500 images serve as the training set to train the neural network, and the rest 500 images are used for verifying the accuracy and the success rate of neural network identification. Then a tag file is generated. Generating self train.txt, val.txt and test.txt according to the trained and tested pictures, wherein the files contain network labels, and the labels are data for labeling a data set and are convenient for training. The image is changed into an image with the size of 256 multiplied by 256 by resize, then a corresponding lmdb file is generated, and training can be started after the mean value of the image is calculated.
The trained neural network weight file and other configuration files are applied to a detection program, and after a detection instruction is sent, the program obtains a real-time data set from an appointed file directory in real time so as to conveniently detect the target crack in real time. And continuously iterating to obtain a cost loss function through forward channel operation and backward propagation operation. After the operations of continuously updating the w and b values and the like by using a gradient descent method, a crack identification diagram with higher precision after feature extraction processing is finally obtained, the process of crack extraction and identification is completed, and the purpose of real-time detection is achieved. The results of the detection are shown in FIGS. 4 and 5.
Based on the bridge defect detection method, the invention also correspondingly provides bridge defect detection equipment, which comprises the following steps: a processor and a memory;
the memory has stored thereon a computer readable program executable by the processor;
the processor, when executing the computer readable program, implements the steps in the bridge defect detecting method according to the above embodiments.
Since the bridge defect detection method has been described in detail above, it is not described herein again.
Based on the bridge defect detection method, the invention further provides a computer-readable storage medium, where one or more programs are stored, and the one or more programs can be executed by one or more processors to implement the steps in the bridge defect detection method according to the embodiments.
Since the bridge defect detection method has been described in detail above, it is not described herein again.
Based on the bridge defect detection method and the bridge defect detection equipment, the embodiment of the invention also correspondingly provides a bridge defect detection system, which comprises the bridge defect detection equipment, the unmanned aerial vehicle, the repair vehicle and a plurality of light sources, wherein the unmanned aerial vehicle is electrically connected with the bridge defect detection equipment and is used for acquiring bridge image data; the repairing vehicle is electrically connected with the bridge defect detection equipment and is used for overhauling according to the detection result of the bridge defect detection equipment; the light sources are respectively located on the unmanned aerial vehicle and the repair vehicle and used for light supplement.
In a preferred embodiment, the bridge defect detection system further comprises sensors, such as levels, gyroscopes, position sensors, etc. The level meter, the gyroscope and the like are used for collecting information of the settlement of the bridge deck, and the comprehensive data is used for judging the damage degree of the bridge deck. And the positions of the defects can be collected by using position sensors such as a Beidou dual-mode positioning module and the like.
Since the bridge defect detection method and apparatus have been described in detail above, no further description is given here.
In summary, the bridge defect detection method, the bridge defect detection device, the bridge defect detection system and the bridge defect detection storage medium provided by the invention have the advantages that the image is shot by the unmanned aerial vehicle, then the image is preprocessed, the peeling defect is detected by the edge detection method, the crack defect is detected by the neural network model, the detection precision is higher, the working efficiency is high, the working period is short, and the safety risk is not generated.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (10)
1. A bridge defect detection method is characterized by comprising the following steps:
acquiring bridge image data acquired by an unmanned aerial vehicle;
preprocessing the bridge image data to obtain a clear image with a background except for the defect removed;
performing edge detection on the image obtained after pretreatment to detect the area of the bridge body with peeling defects;
and processing the preprocessed image by adopting a preset convolutional neural network model so as to detect the position of the crack defect of the bridge body and the size of the crack defect.
2. The bridge defect detection method of claim 1, wherein the bridge body image data at least comprises bridge deck image data, bridge frame image data and pier image data.
3. The bridge defect detection method according to claim 1, wherein the step of preprocessing the bridge image data to obtain a clear image with a background except for the defect removed specifically comprises:
performing framing processing on the bridge image data to obtain a plurality of key frame images;
and processing each key frame image by adopting an image processing technology to obtain a plurality of clear images with the background except the defects removed.
4. The bridge defect detecting method according to claim 3, wherein the step of performing framing processing on the bridge body image data to obtain a plurality of key frame images specifically comprises:
acquiring the total frame number of the bridge image data, reading the bridge image data frame by frame, and taking an image with the number of the frame number being N times of a preset value as a key frame image; wherein N is a natural number not less than 1.
5. The bridge defect detecting method according to claim 3, wherein the step of processing each key frame image by using an image processing technique to obtain a plurality of clear images with backgrounds except for defects removed specifically comprises:
performing graying processing on the key frame image, and converting an image in an RGB domain into an image in a GRY domain;
equalizing the gray level of the image subjected to gray level processing through a histogram;
performing median filtering on the equalized image by adopting a median filtering method;
performing edge enhancement processing on the filtered image by adopting a Laplace algorithm;
and performing background segmentation on the image subjected to edge enhancement processing by adopting a threshold segmentation method to obtain a clear image with the background except the defect removed.
6. The bridge defect detection method according to claim 1, wherein the step of performing edge detection on the preprocessed image to detect the area of the bridge body with the peeling defect specifically comprises:
adopting a Canny edge detection algorithm to carry out edge extraction on the image, and searching a maximum connected domain of the edge after the edge is obtained;
and calculating the area of the maximum connected domain, comparing the area of the maximum connected domain with a preset area threshold, and judging that the maximum connected domain is a region needing to be repaired when the area of the maximum connected domain is larger than the preset area threshold.
7. The bridge defect detecting method of claim 1, wherein the preset convolutional neural network model is a network model based on a YOLOV5 neural network framework.
8. A bridge defect detection device, comprising: a processor and a memory;
the memory has stored thereon a computer readable program executable by the processor;
the processor, when executing the computer readable program, implements the steps in the bridge defect detection method of any one of claims 1-7.
9. A computer-readable storage medium storing one or more programs which are executable by one or more processors to perform the steps of the bridge defect detection method of any one of claims 1-7.
10. A bridge defect detection system comprising the bridge defect detection apparatus of claim 8, further comprising:
the unmanned aerial vehicle is electrically connected with the bridge defect detection equipment and is used for acquiring bridge body image data;
the repairing vehicle is electrically connected with the bridge defect detection equipment and is used for overhauling according to the detection result of the bridge defect detection equipment;
and the light sources are respectively positioned on the unmanned aerial vehicle and the repairing vehicle and used for light supplement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110096158.3A CN112801972A (en) | 2021-01-25 | 2021-01-25 | Bridge defect detection method, device, system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110096158.3A CN112801972A (en) | 2021-01-25 | 2021-01-25 | Bridge defect detection method, device, system and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112801972A true CN112801972A (en) | 2021-05-14 |
Family
ID=75811554
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110096158.3A Pending CN112801972A (en) | 2021-01-25 | 2021-01-25 | Bridge defect detection method, device, system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112801972A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113256601A (en) * | 2021-06-10 | 2021-08-13 | 北方民族大学 | Pavement disease detection method and system |
| CN113533368A (en) * | 2021-05-24 | 2021-10-22 | 国网河北省电力有限公司衡水供电分公司 | Electric pole crack detection device, control method and control terminal |
| CN113640307A (en) * | 2021-08-31 | 2021-11-12 | 郑州铁路职业技术学院 | Track condition monitoring method adopting machine vision |
| CN113658095A (en) * | 2021-07-09 | 2021-11-16 | 浙江大学 | Engineering pattern review identification processing method and device for drawing of manual instrument |
| CN114511568A (en) * | 2022-04-20 | 2022-05-17 | 西安博康硕达网络科技有限公司 | Expressway bridge overhauling method based on unmanned aerial vehicle |
| CN114851223A (en) * | 2022-05-24 | 2022-08-05 | 武汉理工大学 | Bionic robot for wall detection, image processing device and working method |
| CN115953366A (en) * | 2022-12-14 | 2023-04-11 | 广州市斯睿特智能科技有限公司 | Weld joint detection method, system and device based on reflection image and storage medium |
| CN115995056A (en) * | 2023-03-22 | 2023-04-21 | 南京航空航天大学 | Automatic bridge disease identification method based on deep learning |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050140787A1 (en) * | 2003-11-21 | 2005-06-30 | Michael Kaplinsky | High resolution network video camera with massively parallel implementation of image processing, compression and network server |
| US20070071304A1 (en) * | 2005-09-27 | 2007-03-29 | Sharp Kabushiki Kaisha | Defect detecting device, image sensor device, image sensor module, image processing device, digital image quality tester, and defect detecting method |
| CN101153850A (en) * | 2006-09-30 | 2008-04-02 | 长安大学 | Method and system for detecting asphalt mixture |
| CN105448041A (en) * | 2016-01-22 | 2016-03-30 | 苏州望湖房地产开发有限公司 | A human body falling intelligent control system and method |
| CN105975972A (en) * | 2016-04-27 | 2016-09-28 | 湖南桥康智能科技有限公司 | Bridge crack detection and characteristic extraction method based on image |
| CN106846316A (en) * | 2017-02-10 | 2017-06-13 | 云南电网有限责任公司电力科学研究院 | A kind of GIS inside typical defect automatic distinguishing method for image |
| US20180253836A1 (en) * | 2015-06-16 | 2018-09-06 | South China University Of Technology | Method for automated detection of defects in cast wheel products |
| WO2018214195A1 (en) * | 2017-05-25 | 2018-11-29 | 中国矿业大学 | Remote sensing imaging bridge detection method based on convolutional neural network |
| CN109444171A (en) * | 2018-09-18 | 2019-03-08 | 山东理工大学 | Integrated Bridges Detection based on unmanned plane |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN110378879A (en) * | 2019-06-26 | 2019-10-25 | 杭州电子科技大学 | A kind of Bridge Crack detection method |
| US20190378261A1 (en) * | 2018-06-11 | 2019-12-12 | Dynatek Labs, Inc. | Automated failure detection for medical device testing systems and methods |
| CN110648349A (en) * | 2019-09-05 | 2020-01-03 | 南开大学 | Weld defect segmentation method based on background subtraction and connected region algorithm |
| CN112233092A (en) * | 2020-10-16 | 2021-01-15 | 广东技术师范大学 | Deep learning method for intelligent defect detection of unmanned aerial vehicle power inspection |
-
2021
- 2021-01-25 CN CN202110096158.3A patent/CN112801972A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050140787A1 (en) * | 2003-11-21 | 2005-06-30 | Michael Kaplinsky | High resolution network video camera with massively parallel implementation of image processing, compression and network server |
| US20070071304A1 (en) * | 2005-09-27 | 2007-03-29 | Sharp Kabushiki Kaisha | Defect detecting device, image sensor device, image sensor module, image processing device, digital image quality tester, and defect detecting method |
| CN101153850A (en) * | 2006-09-30 | 2008-04-02 | 长安大学 | Method and system for detecting asphalt mixture |
| US20180253836A1 (en) * | 2015-06-16 | 2018-09-06 | South China University Of Technology | Method for automated detection of defects in cast wheel products |
| CN105448041A (en) * | 2016-01-22 | 2016-03-30 | 苏州望湖房地产开发有限公司 | A human body falling intelligent control system and method |
| CN105975972A (en) * | 2016-04-27 | 2016-09-28 | 湖南桥康智能科技有限公司 | Bridge crack detection and characteristic extraction method based on image |
| CN106846316A (en) * | 2017-02-10 | 2017-06-13 | 云南电网有限责任公司电力科学研究院 | A kind of GIS inside typical defect automatic distinguishing method for image |
| WO2018214195A1 (en) * | 2017-05-25 | 2018-11-29 | 中国矿业大学 | Remote sensing imaging bridge detection method based on convolutional neural network |
| US20190378261A1 (en) * | 2018-06-11 | 2019-12-12 | Dynatek Labs, Inc. | Automated failure detection for medical device testing systems and methods |
| CN109444171A (en) * | 2018-09-18 | 2019-03-08 | 山东理工大学 | Integrated Bridges Detection based on unmanned plane |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN110378879A (en) * | 2019-06-26 | 2019-10-25 | 杭州电子科技大学 | A kind of Bridge Crack detection method |
| CN110648349A (en) * | 2019-09-05 | 2020-01-03 | 南开大学 | Weld defect segmentation method based on background subtraction and connected region algorithm |
| CN112233092A (en) * | 2020-10-16 | 2021-01-15 | 广东技术师范大学 | Deep learning method for intelligent defect detection of unmanned aerial vehicle power inspection |
Non-Patent Citations (3)
| Title |
|---|
| 刘小燕: "基于数字图像处理的混凝土桥梁底面裂缝的检测", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》, no. 04, 15 April 2015 (2015-04-15), pages 2 * |
| 闫龙川，高德荃，李君婷译；迪潘简•萨卡尔: "《数据科学与工程技术丛书 Python文本分析 第2版》", 31 October 2020, pages: 390 * |
| 陈志军作: "《智能网联环境下车辆运动行为理解方法》", 31 December 2020, pages: 56 - 57 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113533368A (en) * | 2021-05-24 | 2021-10-22 | 国网河北省电力有限公司衡水供电分公司 | Electric pole crack detection device, control method and control terminal |
| CN113533368B (en) * | 2021-05-24 | 2022-10-21 | 国网河北省电力有限公司衡水供电分公司 | Electric pole crack detection device, control method and control terminal |
| CN113256601A (en) * | 2021-06-10 | 2021-08-13 | 北方民族大学 | Pavement disease detection method and system |
| CN113658095A (en) * | 2021-07-09 | 2021-11-16 | 浙江大学 | Engineering pattern review identification processing method and device for drawing of manual instrument |
| CN113640307A (en) * | 2021-08-31 | 2021-11-12 | 郑州铁路职业技术学院 | Track condition monitoring method adopting machine vision |
| CN113640307B (en) * | 2021-08-31 | 2023-10-10 | 郑州铁路职业技术学院 | Rail condition monitoring method adopting machine vision |
| CN114511568A (en) * | 2022-04-20 | 2022-05-17 | 西安博康硕达网络科技有限公司 | Expressway bridge overhauling method based on unmanned aerial vehicle |
| CN114511568B (en) * | 2022-04-20 | 2022-07-22 | 西安博康硕达网络科技有限公司 | Expressway bridge overhauling method based on unmanned aerial vehicle |
| CN114851223A (en) * | 2022-05-24 | 2022-08-05 | 武汉理工大学 | Bionic robot for wall detection, image processing device and working method |
| CN115953366A (en) * | 2022-12-14 | 2023-04-11 | 广州市斯睿特智能科技有限公司 | Weld joint detection method, system and device based on reflection image and storage medium |
| CN115995056A (en) * | 2023-03-22 | 2023-04-21 | 南京航空航天大学 | Automatic bridge disease identification method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112801972A (en) | Bridge defect detection method, device, system and storage medium | |
| US20210319561A1 (en) | Image segmentation method and system for pavement disease based on deep learning | |
| CN111402209B (en) | U-Net-based high-speed railway steel rail damage detection method | |
| CN114693615A (en) | Deep learning concrete bridge crack real-time detection method based on domain adaptation | |
| CN110473221B (en) | Automatic target object scanning system and method | |
| CN111489352A (en) | Tunnel gap detection and measurement method and device based on digital image processing | |
| CN113592861A (en) | Bridge crack detection method based on dynamic threshold | |
| CN110207592A (en) | Building cracks measurement method, device, computer equipment and storage medium | |
| CN109489724A (en) | A kind of tunnel safe train operation environment comprehensive detection device and detection method | |
| CN110599469A (en) | Method and system for detecting defects of key parts of motor train unit and electronic equipment | |
| CN113240623B (en) | Pavement disease detection method and device | |
| CN111080600A (en) | Fault identification method for split pin on spring supporting plate of railway wagon | |
| CN111080613B (en) | Image recognition method for damage fault of wagon bathtub | |
| CN114119505A (en) | Method and device for detecting chip adhesion area defects | |
| CN116485779B (en) | Adaptive wafer defect detection method and device, electronic equipment and storage medium | |
| CN114926407A (en) | Steel surface defect detection system based on deep learning | |
| CN112102201A (en) | Image shadow reflection eliminating method and device, computer equipment and storage medium | |
| CN110728269B (en) | High-speed rail contact net support pole number plate identification method based on C2 detection data | |
| CN114419421A (en) | Subway tunnel crack identification system and method based on images | |
| CN117422699A (en) | Highway detection method, highway detection device, computer equipment and storage medium | |
| CN111325724B (en) | Tunnel crack region detection method and device | |
| CN113762247A (en) | Road crack automatic detection method based on significant instance segmentation algorithm | |
| CN116797602A (en) | Surface defect identification method and device for industrial product detection | |
| CN112288726A (en) | Method for detecting foreign matters on belt surface of underground belt conveyor | |
| CN116363655A (en) | Financial bill identification method and system |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210514 |