Crack Detection in Construction Field based on Improved Faster R-CNN Algorithm
Driven by artificial intelligence technology,the intelligentization of national building facilities is also developing rapidly,and the problem of wall crack detection is also receiving more and more attention.Aiming at the problem of low accuracy of traditional manual detection of building wall cracks,this paper proposes an improved Faster R-CNN algorithm for wall crack detection.Firstly,the experimental dataset was made and enhanced,then the ResNet50 residual network was used to replace the VGG16 network module for feature extraction,then the FPN feature pyramid module was added to improve the multi-scale detection ability of the model,and finally,the EIoU loss function was used to improve the accuracy of the model detection.The experiments showed that the improved algorithm in this paper has greatly improved the detection ability,and the mAP value reaches 93.5％,which can meet the demand of high-precision detection.
万方数据
维普
