Surface Crack Detection Method for Coal Rock Based on Improved YOLOv5

Abstract

Coal mine safety may be able to be ensured via the real-time identification of cracks in rock and coal surfaces. Traditional crack identification methods have the disadvantages of slow speed and low precision. This work suggests an improved You Only Look Once version 5 (YOLOv5) detection model. In this study, we improved YOLOv5 from the perspective of three aspects: a Ghost module was introduced into the backbone network to lighten the model; a coordinate attention mechanism was added; and ECIOU_Loss is proposed as a loss function in this paper to achieve the co-optimization of crack detection speed and accuracy and to meet the deployment requirements in the embedded terminal. The results demonstrate that the improved YOLOv5 has a 92.8% mean average precision (mAP) with an 8 MB model size, and the speed of recognition was 103 frames per second. Compared to the original method, there was a 53.4% reduction in the number of parameters, a detection speed that was 1.9 times faster, and a 1.7% improvement in the mAP. The improved YOLOv5 can effectively locate cracks in real time and offers a new technique for the early warning of coal and rock dynamic hazards.

Research on real-time perception method of key targets in the comprehensive excavation working face of coal mine

“…Therefore, to solve the above problems, this study adopted the ECIoU loss function [19], which combines the advantages of both CIoU and EIoU. The ECIoU loss function utilizes CIoU to change the aspect ratio of the prediction frame until the function converges to the appropriate range and then utilizes EIoU to adjust each edge of the prediction frame until the function converges to the correct value.…”

YOLO-UAV: Object Detection Method of Unmanned Aerial Vehicle Imagery Based on Efficient Multi-Scale Feature Fusion

“…Here, ρ 2 (h gt ,h) When dealing with a box with a distant edge, the computation of EIoU can slow down and convergence isn't achieved in advance. To resolve this matter, Chen et al [44] proposed the ECIoU (Efficient Complete-IoU), that is able to improve the adjustment of the predicted box and speed up its regression convergence rate. The ECIoU is based on the combination of CIoU and EIoU loss functions.…”

Intelligent Detection of Road Cracks Based on Improved YOLOv5

“…Among them, YOLOv5 has achieved smaller model size and lower computing resource consumption by using modern deep learning technology and structural optimization. It shows high speed and precision in real-time target detection tasks, and is very friendly to practical engineering applications [6]. Therefore, based on YOLOv5, this paper proposes a model combining Swin Transformer and BiFPN model, which Zhiyan Zhou et al…”