Research on intelligent detection of cables of raw rice bridge based on rope climbing robot

In order to realize the accurate recognition of bridge cable surface damage, an image recognition method based on improved YOLOV8 model is proposed. The image information of the bridge cable surface is collected by the rope climbing robot, which is an intelligent detection device for bridge cables, and then the image data set with defects on the cable surface is selected. Firstly, Diverse Branch Block (DBB) was introduced to construct a C2fDBB module to replace the C2f module in the original algorithm and enhance the capability of multi-scale feature extraction. Secondly, in the original YOLOV8 model, the MSCA attention module was combined with the common CBAM attention module, and the MCASAM attention module was innovatively designed and added to the last layer of the feature extraction network, which improved the multi-scale feature capture ability of the model and the detection sensitivity of small targets. The experimental results in the homemade cable defect dataset show that the average accuracy of the improved algorithm mAP50 is 0.9357 and the average detection time is 0.176 compared with the original algorithm, which are 6.97% and 6.82%, respectively. The experimental results show that the improved YOLOv8 model has better performance in defect detection.