Research on Intelligent Defect Detection Method for Urban Drainage Pipelines
To address the problems of low automatic detection accuracy and inaccurate target positioning for urban drainage pipe-line defects,an improved YOLOv8 drainage pipeline defect detection model is proposed.This model introduces receptive field attention convolution to the baseline model and constructs the C2F receptive field attention convolution(RFAConv)module,enhancing the model's ability to extract defect features through interactive adaptive learning of spatial receptive fields and convolutions.Additional-ly,a hybrid attention high-order and low-order feature fusion network is proposed,which effectively fuses the low-order and high-or-der features of three different scales by the backbone and neck outputs,enhancing the models ability to learn the global contextual in-formation of the image.The Inner-MPDIoU loss function is designed by comprehensively analyzing the factors affecting the bounding box regression,such as overlap,distance between the center points,and deviation in width and height.The model is suitable for the defect detection of different sizes,and improves the positioning accuracy of the defect target boundary box.Experimental results show that compared with the baseline model,the improved model achieves an average detection accuracy of 93.9％,with an increase of 3.7％,the missed detection rate and false detection rate of the improved model are reduced to 9.1％and 17.6％,with decreases of 3.2％and 2.7％,respectively.
万方数据
维普
