Three-Stage Pavement Crack Localization and Segmentation Algorithm Based on Digital Image Processing and Deep Learning Techniques

Abstract

The image of expressway asphalt pavement crack disease obtained by a three-dimensional line scan laser is easily affected by external factors such as uneven illumination distribution, environmental noise, occlusion shadow, and foreign bodies on the pavement. To locate and extract cracks accurately and efficiently, this article proposes a three-stage asphalt pavement crack location and segmentation method based on traditional digital image processing technology and deep learning methods. In the first stage of this method, the guided filtering and Retinex methods are used to preprocess the asphalt pavement crack image. The processed image removes redundant noise information and improves the brightness. At the information entropy level, it is 63% higher than the unpreprocessed image. In the second stage, the newly proposed YOLO-SAMT target detection model is used to locate the crack diseases in asphalt pavement. The model is 5.42 percentage points higher than the original YOLOv7 model on mAP@0.5, which enhances the recognition and location ability of crack diseases and reduces the calculation amount for the extraction of crack contour in the next stage. In the third stage, the improved k-means clustering algorithm is used to extract cracks. Compared with the traditional k-means clustering algorithm, this method improves the accuracy by 7.34 percentage points, the true rate by 6.57 percentage points, and the false positive rate by 18.32 percentage points to better extract the crack contour. To sum up, the method proposed in this article improves the quality of the pavement disease image, enhances the ability to identify and locate cracks, reduces the amount of calculation, improves the accuracy of crack contour extraction, and provides a new solution for highway crack inspection.

Pavement Defect Detection Algorithm Based on Improved YOLOv7 Complex Background

“…And because the calculation of GELU is relatively complex, replacing too many convolution modules will consume more computing resources, thus reducing the detection speed of the model. Therefore, this study uses the GELU [23] GELU is an activation function based on the Gaussian error function, the full name is "Gaussian Error Linear Unit". As an excellent activation function proposed in 2020, compared with SiLU, the function will not treat all x less than or equal to 0 equally as 0. , taking all 0 will cause the derivative to be equal to 0, causing the gradient to disappear, so the GELU activation function eliminates the problem of the gradient disappearing.…”

Particle size estimation of industrial raw materials based on improved YOLOv7

“…Aiming at the problems of smaller and denser industrial raw material particles, large or skewed prediction frame after training, slower speed when detecting dense small targets, and lower detection accuracy, the following three improvements are made: (1). Add segmentation [15] segmentation module to automatically generate the detection frame according to the segmentation. Adding this module to the YOLOv7 network can achieve detection and segmentation of the target by segmenting the dataset, so that the prediction frame fits the target better.…”

Assessing the performance of YOLOv5, YOLOv6, and YOLOv7 in road defect detection and classification: a comparative study

“…Next, Aburaed et al [13] evaluated the performance of YOLOv6 compared to YOLOv5 on detecting craters, where the claims that YOLOv6 would outperform YOLOv5 still can't be proven as their performance was inconsistence in every scenario. Meanwhile, Yang et al [14] proposed a three-stage crack location and segmentation method where it is first filtered by the Retinex method to remove redundant noise, followed by detection process where YOLO-SAMT was introduced, and lastly processed by K-means clustering to extract the cracks. YOLO-SAMT is an enhanced algorithm where YOLOv7 architecture is integrated with SimAM and transformer, which shows a 5.42% higher mAP score than the original YOLOv7.…”