Intelligent Classification Method for Subgrade Disease Based on Improved GoogleNet-ResNet Algorithm
There are the shortages of complex disease identification and poor utilization of multi-view radar image features in sub-grade disease classification algorithms,an intelligent classification method for roadbed diseases based on an improved GoogleNet-Res-Net algorithm is proposed.Firstly,the coordinate attention and improved Inception modules are introduced to optimize the GoogleNet network structure.Then,the improved GoogleNet is utilized to learn the c-scan data features and eliminate non-target diseases,a-chieving the coarse classification of the disease targets.Finally,the b-scan data classified as the diseases is input into the ResNet50 model based on the transfer learning to achieve the fine classification of the diseases.The results show that the accuracy of the im-proved GoogleNet reaches by 98.2％for coarse disease classification,and the detection speed by 90.9 fps.The accuracy of the disease sub-classification of ResNet50 based on the transfer learning reaches by 90.5％,and the detection speed by 52.6 fps.The accuracy of the proposed algorithm is 10.1％higher than that of the improved GoogleNet network,and 7.4％higher than that of the ResNet50 network.This algorithm effectively improves the recognition accuracy and efficiency of roadbed disease detection.
NETL
NSTL
万方数据
