Research On Transmission Line Small Target Recognition Algorithm Based On YOLO

With the development of society and the progress of science and technology,and researchers’ research on deep learning is becoming more and more thorough,making object detection based on deep learning more and more extensive.In this paper,YOLO algorithm is mainly introduced into the identification and positioning of small targets in transmission lines,so as to achieve the requirements of real-time detection and accurate identification and positioning of targets.The specific work of this paper is as follows:(1)To meet the requirements of the YOLO algorithm,complete datasets of shock hammers,spacers and insulators were built.(2)In order to improve the speed and accuracy of small target recognition of transmission lines,this paper first analyzes the advantages and disadvantages of traditional YOLOv3 in small target detection and recognition of transmission lines,and focuses on the analysis of the loss function and spatial pyramid module that affect the accuracy and speed of small target recognition.The analysis results show that YOLOv3 with improved loss function and added space pyramid pooling module can meet the accuracy of small target recognition of transmission lines,but the detection speed is slow.(3)In order to solve the shortcomings of the improved YOLOv3 detection rate,this paper proposes an improved YOLOv5 s,which uses the traditional YOLOv5 s model as the basic network framework for training to improve the YOLOv5 s model: First,by adding the CBAM attention mechanism,the two attention frames of channel and space are fused to obtain richer detailed features,which is conducive to the recognition of small targets and changes the performance of the model;Second,the FRe LU activation function is used instead of the Re LU activation function to add simple and effective spatial conditions,which does not affect the running speed and realizes the efficient extraction of the detection target features.By comparing the detection results of the improved YOLOv5 s and the improved YOLOv3 algorithm on the small target dataset of transmission lines,it can be analyzed that the improved YOLOv5 s not only greatly improves the detection accuracy,but also takes into account the target requirements of real-time detection.This article for the study of the transmission line of small target image recognition,provides a new solution,by comparing the improved YOLOv5 s and other common target detection algorithm in small target recognition test results on the transmission line,can analyze the improved YOLOv5 s in solving transmission line on the accuracy and speed of the small target recognition has made a lot of ascension,It is more suitable for transmission line target detection.Figure [63] table [10] reference [84]...