Research And Application Of Object Detection Methods On Power Cable Inspection Pictures Based On Convolutional Neural Network

With the popularization and application of digital technology in power cable inspection,the continuous accumulation of data puts forward higher requirements on the efficiency of data processing.In particular,the information of sensitive objects contained in the inspection pictures plays an important role in preventing external damage.At present,the processing of inspection pictures relies on labor,not only the processing efficiency is not high,but also due to the quality of personnel and the responsibility of the problem,the phenomenon of missed detection sometimes occurs.The object detection technology based on convolutional neural network can effectively reduce the workload of image data processing,and provides an effective solution for the development of the inspection picture intelligent module.This paper investigated the development status of intelligent cable inspection,image classification,object detection,etc.The basic structure and key technologies of object detection model are analyzed with Faster R-CNN,SSD and YOLOv3,and data enhancement preprocessing and model are obtained.Structural adjustment and adjustment of model component parameters can improve detection speed and accuracy.This paper proposes that the problem of object detection,object detection host and monitoring platform fusion on a specific data set needs to be solved,and a YOLOV3-based cable inspection image object detection model YOLOv3 PCIP(YOLOv3 for power cable inspection picture)is established.The original model has been improved to varying degrees and meets engineering application requirements.Finally,using Python and JAVA mixed programming,spring framework development and deployment,embedded web container and timing tasks,the object detection independent host and monitoring platform server integration,to achieve industrial production applications.In this paper,the artificial intelligence method represented by deep learning convolutional neural network is introduced in the field of cable inspection image processing,which enriches the processing methods of cable inspection data,and provides a referencefor other data processing fields such as natural language processing,cable inspection data system architecture and so on.The proposed object detection model YOLOv3 PCIP is applied to the Nanjing cable intelligent monitoring platform,which solves the problem of low efficiency and high error rate of manual image inspection.The effectiveness of the object detection model is verified in practical work.