Detection and recognition of digital instrument in substation using improved YOLO-v3

In order to monitor substation intelligently, it is of significance to obtain substation instrument automatically and accurately. This paper adopts the digital instrument of the substation in the actual scene as the research object and proposes a detection and identification method based on the improved YOLO-v3 for the substation digital instrument. In order to enrich the limited image data, this paper augments the specific image data of the number of substations collected and constructs the data set. Based on YOLO-v3, aiming at the problem of the accuracy of substation instrument detection and identification, and considering the real-time performance comprehensively, this pager proposes an improved YOLO-v3 model by using PANet structure. The effectiveness of the proposed method is verified according to the substation digital instrument detection experiment. Experimental results show that the improved YOLO-v3 is 0.23% higher than the classical YOLO-v3 network concerning mean average precision, and it has better accuracy in substation digital instrument detection and identification. The proposed method can still guarantee a real-time performance, and the detection frames per second (FPS) of image processing is 29 f/s; it meets the actual substation intelligent data acquisition, detection and identification engineering needs.

Automated summary

This paper proposes a detection and identification method based on the improved YOLO-v3 for substation digital instruments, aiming to monitor substations intelligently. By augmenting the image dataset and using the PANet structure, the proposed method achieves accurate detection and identification of substation instruments with real-time performance, meeting the actual needs of substation data acquisition and engineering.