Retraction Note to: Power Equipment Defect Detection Algorithm Based on Deep Learning

Abstract

Ensuring the safety of power equipment has an important meaning to the development of the country and people’s lives. The safety of power equipment cannot be separated from the timely detection of power equipment defects. The current mainstream power equipment defect detection method is to take a picture of the power equipment in a complex environment through a drone and manually select from the video. This method requires a great deal of manpower and resources. This paper intends to use a target detection algorithm, which can replace the manual detection of power equipment defects based on UAV pictures. Most of the traditional power equipment defect detection algorithms can only target specific scenarios and do not have strong generalization. This paper compares the current target detection algorithm commonly used in industry, investigates its structure, performance, advantages and disadvantages, and its performance in public data, and summarizes the basic target detection network architecture which is most suitable for power equipment defects. In this paper, a new target detection algorithm is used to self-learn through deep neural networks to perform defect inspection of power equipment in UAV pictures with different backgrounds, illuminations, and scales. Through the reasonable selection of the target detection algorithm framework and the improvement of multiple steps, the proposed algorithm for detecting power equipment defects is improved by nearly 60% compared with the conventional deep learning target detection algorithm.