Design And Implementation Of ACCC Carbon Core Defect Imaging And Automatic Inspection System

The aluminum conductor composite core(ACCC)plays an important role in increasing the transmission capacity of transmission lines.However,due to shortcomings such as inability to withstand bending,frequent disconnection accidents occur.At the same time,the extremely long line brings greater The workload is easy to cause fatigue of the inspectors and lead to missed inspections.Therefore,automatic inspection of carbon core defects of ACCC wires is of great significance.Aiming at the problems of heavy manual flaw detection,cumbersome defect detection operations,and insufficient defect detection stability,this thesis designs and implements an ACCC carbon core defect imaging and automatic detection system,which uses deep learning technology to implement algorithm modules and data management Module,interaction and display module,communication module,report module,among them the algorithm module realizes the wire image standardization,automatic defect recognition,network training function,the data management module realizes the ACCC wire data annotation,storage and expansion function,the interaction and display module The visualization and image enhancement functions of ACCC wire images are realized,the communication module realizes the communication function between the upper computer and the lower computer,and the report log module realizes the system data log generation function.This system meets the requirements of ACCC carbon core defect detection automation,fast detection speed,and stable detection rate,improves the efficiency and accuracy of ACCC flaw detection,and has certain practical value.In the process of implementation,this thesis proposes a standardized preprocessing method for the characteristics of ACCC wire image shape bending and brightness inconsistency.By preprocessing the wire image data for shape standardization and brightness consistency,the system’s performance is effectively improved.Defect detection accuracy;at the same time,in view of the problem of insufficient ACCC data samples,data expansion methods for different types of defects are designed to effectively solve the problem of data proportion imbalance.At the same time,the neural network is trained with expanded data to make it have higher defect detection accuracy.The optimization and improvement of the algorithm have effectively improved the defect detection performance of the system,which has certain theoretical value.