Open AccessResearch on the use of YOLOv3 image processing algorithm in power plants
Author(s) -
Junjie Zhao
Publication year - 2019
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/677/4/042059
Subject(s) - computer science , image processing , image (mathematics) , power (physics) , artificial intelligence , computer vision , digital image processing , algorithm , residual , physics , quantum mechanics
The power system is a guarantee system for nation-building, and it is inseparable from the use of electricity everywhere in life. There are many links in the production of power plants that need to be well monitored, but good real-time monitoring has always been a problem. The YOLOv3 algorithm draws on the residual network structure in the target detection of image processing, forms a deeper network level, and multi-scale detection, which improves the detection of mAP and small objects. Applying the YOLOv3 algorithm to the image processing of power plants can effectively improve the efficiency of image detection in power plants. Based on the different network structures of YOLOv3, this paper studies the power plant image processing in terms of detection speed and accuracy. The research shows that although the results of different network structures are different, the application of YOLOv3 in power plant image processing is still at the detection speed. Fast, and its detection accuracy is also very high.