Pavement crack identification method, device and system, and storage medium

Abstract

The invention discloses a pavement crack identification method, device and system, and a storage medium. The method comprises the following steps: S1, acquiring a historical pavement image; s2, carrying out disease marking on the historical pavement image to obtain pavement crack condition statistical information; s3, obtaining a convolutional neural network model for pavement crack recognition according to the pavement crack condition statistical information; and S4, inputting a to-be-processed pavement image into the convolutional neural network model to realize recognition of pavement diseases. According to the technical scheme, the problems that an existing method is low in recognition efficiency and low in target detection accuracy are solved, the crack condition of the asphalt pavement is effectively detected, and real-time data are provided for pavement crack maintenance.

Description

Technical Field

[0001] The present invention belongs to the technical field of crack detection, and in particular relates to a method and device, a system, and a storage medium for identifying road cracks. Background Art

[0002] As one of the main forms of urban transportation construction, asphalt pavement plays a key role in urban economic development and social life. However, due to the special use environment and its own properties, asphalt pavement is prone to cracks, fatigue, deformation, etc. during use, which seriously affects the service life and safety of the road.

[0003] With the acceleration of urbanization and the increasing traffic congestion, traditional road detection methods can no longer meet the needs of modern traffic management. Road automation detection technology has emerged. It uses an intelligent system to achieve fully automated detection of road construction, maintenance and management, providing real-time data support for urban traffic management. Traditional asphalt pavement crack detection mainly relies on manual methods, which have problems such as low efficiency, high cost, and inaccurate detection results. With the continuous increase in traffic flow and the extension of road service life, this manual detection method can no longer meet the needs of modern highway maintenance and management. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and device, system and storage medium for pavement crack identification, which solves the problems of low recognition efficiency and low detection target accuracy of existing methods, effectively detects cracks in asphalt pavement, and provides real-time data for pavement crack maintenance.

[0005] To achieve the above object, the present invention adopts the following technical solution:

[0006] A method for identifying pavement cracks, comprising:

[0007] Step S1, obtaining historical road surface images;

[0008] Step S2, marking the defects of historical road surface images to obtain statistical information on road surface cracks;

[0009] Step S3, obtaining a convolutional neural network model for pavement crack identification based on pavement crack situation statistical information;

[0010] Step S4: input the road surface image to be processed into the convolutional neural network model to identify road surface defects.

[0011] Preferably, step S3 comprises:

[0012] Normalize the statistical information of pavement cracks;

[0013] The yolov8 model is trained according to the normalized statistical information of pavement cracks to obtain a convolutional neural network model for pavement crack identification.

[0014] Preferably, in step S4, non-maximum suppression is used to screen the crack bounding box output by the convolutional neural network model.

[0015] The present invention also provides a road surface crack identification device, comprising:

[0016] An acquisition module, used to acquire historical road surface images;

[0017] The annotation module is used to annotate the defects of historical pavement images and obtain the statistical information of pavement cracks;

[0018] A processing module, used to obtain a convolutional neural network model for pavement crack identification based on pavement crack situation statistical information;

[0019] The recognition module is used to input the road surface image to be processed into the convolutional neural network model to identify road surface defects.

[0020] Preferably, the processing module comprises:

[0021] A normalization unit, used for normalizing the statistical information of pavement cracks;

[0022] The training unit is used to train the yolov8 model according to the normalized road crack situation statistical information to obtain a convolutional neural network model for road crack identification.

[0023] Preferably, the recognition module uses non-maximum suppression to screen the crack bounding box output by the convolutional neural network model.

[0024] The present invention also provides a road surface crack identification system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a road surface crack identification method when executed by the processor.

[0025] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the pavement crack identification method when running.

[0026] The present invention establishes a convolutional neural network model for pavement crack recognition based on yolov8, uses the convolutional neural network model to recognize pavement disease images, and uses non-maximum suppression to filter the crack boundary box output by the convolutional neural network model. The technical solution of the present invention effectively solves the problems of low recognition efficiency and low detection target accuracy of the existing methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0028] Figure 1 The present invention is a flow chart of a method for identifying pavement cracks according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0031] Embodiment 1:

[0032] like Figure 1 As shown, an embodiment of the present invention provides a method for identifying pavement cracks, including:

[0033] Step S1, obtaining historical road surface images;

[0034] Step S2, marking the defects of historical road surface images to obtain statistical information on road surface cracks;

[0035] Step S3, obtaining a convolutional neural network model for pavement crack identification based on pavement crack situation statistical information;

[0036] Step S4: input the road surface image to be processed into the convolutional neural network model to identify road surface defects.

[0037] As an implementation of the embodiment of the present invention, step S3 includes:

[0038] Step S31, normalizing the statistical information of road cracks;

[0039] Step S32: train the yolov8 model according to the normalized road crack situation statistical information to obtain a convolutional neural network model for road crack identification.

[0040] Further, in step S31, the data is normalized using the following formula:

[0041]

[0042] Among them, X ’ is the normalized data; X is the statistical information of pavement cracks; X max, X min They are respectively the maximum and minimum values ​​in the statistical information of pavement crack conditions.

[0043] Further, in step S32, a convolutional neural network model is established based on the yolov8 algorithm to determine the following parameters:

[0044] Ephchs: total number of training cycles

[0045] Batch: The number of images in a batch training

[0046] The loss functions include: regression loss function and classification loss function;

[0047] The GIOU loss function is used to measure the loss of the bounding box, as shown in the following formula

[0048] GIOU loss =1-GIOU

[0049]

[0050] Among them, GIOU represents the GIOU distance between the predicted value of the target detection box and the true target box, IOU represents the intersection-over-union ratio between the predicted box and the true box, Au represents the union of the predicted boxes, Bu represents the union of the true boxes, and C represents the area of ​​the minimum closed convex polygon of the predicted box and the true box.

[0051] The DFL loss function is quoted as follows:

[0052] DFL(S i ,S i+1 )=-((y i+1 -y)log(S i )+(yy i )log(S i+1 ))

[0053] Among them, S i , S i+1 The "prediction value" and "near prediction value" of the network output are y, y i ,y i+1 That is, the "actual value" of the label, "label integral value", and "adjacent label integral value".

[0054] BCE (binary cross entropy) is used as the classification loss function, as shown in the following formula:

[0055]

[0056] Among them, y i Indicates the label of sample i, the positive class is 1, the negative class is 0, P i Indicates the probability that sample i is predicted to be a positive class, and the base of log is e

[0057] The activation function is shown in the following formula:

[0058] ReLU(x)=max(x,0)

[0059] Where x is the input value.

[0060] As an implementation of an embodiment of the present invention, in step S4, non-maximum suppression is used to screen the crack boundary box output by the convolutional neural network model.

[0061] Embodiment 2:

[0062] The embodiment of the present invention further provides a road surface crack identification device, comprising:

[0063] An acquisition module, used to acquire historical road surface images;

[0064] The annotation module is used to annotate the defects of historical pavement images and obtain the statistical information of pavement cracks;

[0065] A processing module, used to obtain a convolutional neural network model for pavement crack identification based on pavement crack situation statistical information;

[0066] The recognition module is used to input the road surface image to be processed into the convolutional neural network model to identify road surface defects.

[0067] As an implementation of the embodiment of the present invention, the processing module includes:

[0068] A normalization unit, used for normalizing the statistical information of pavement cracks;

[0069] The training unit is used to train the yolov8 model according to the normalized road crack situation statistical information to obtain a convolutional neural network model for road crack identification.

[0070] As an implementation method of an embodiment of the present invention, the recognition module uses non-maximum suppression to screen the crack boundary box output by the convolutional neural network model.

[0071] Embodiment 3:

[0072] An embodiment of the present invention further provides a pavement crack identification system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a pavement crack identification method when executed by the processor.

[0073] Embodiment 4:

[0074] An embodiment of the present invention further provides a storage medium, on which a computer program is stored, and the computer program executes the pavement crack identification method when running.

[0075] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for identifying pavement cracks, characterized in that: include: Step S1, obtaining historical road surface images; Step S2, marking the defects of historical road surface images to obtain statistical information on road surface cracks; Step S3, obtaining a convolutional neural network model for pavement crack identification based on pavement crack situation statistical information; Step S4: input the road surface image to be processed into the convolutional neural network model to identify road surface defects.

2. The method for identifying pavement cracks according to claim 1, characterized in that: Step S3 includes: Normalize the statistical information of pavement cracks; The yolov8 model is trained according to the normalized statistical information of pavement cracks to obtain a convolutional neural network model for pavement crack identification.

3. The method for identifying pavement cracks according to claim 2, characterized in that: In step S4, non-maximum suppression is used to screen the crack bounding box output by the convolutional neural network model.

4. A road crack identification device, characterized in that: include: An acquisition module, used to acquire historical road surface images; The annotation module is used to annotate the defects of historical pavement images and obtain the statistical information of pavement cracks; A processing module, used to obtain a convolutional neural network model for pavement crack identification based on pavement crack situation statistical information; The recognition module is used to input the road surface image to be processed into the convolutional neural network model to identify road surface defects.

5. The road surface crack identification device according to claim 4, characterized in that: The processing modules include: A normalization unit, used for normalizing the statistical information of pavement cracks; The training unit is used to train the yolov8 model according to the normalized road crack situation statistical information to obtain a convolutional neural network model for road crack identification.

6. The road surface crack identification device according to claim 5, characterized in that: The recognition module uses non-maximum suppression to filter the crack bounding boxes output by the convolutional neural network model.

7. A road crack identification system, characterized in that: include: A memory and a processor, wherein the memory stores a computer program executed by the processor, and when the computer program is executed by the processor, the pavement crack identification method according to any one of claims 1 to 3 is executed.

8. A storage medium, characterized in that: The storage medium stores a computer program, which executes the pavement crack identification method according to any one of claims 1 to 3 when running.

Images