Power Equipment Identification Based on Single Shot Detector

In recent years, more and more deep learning technologies have been extensively used in various fields, such as intelligent medical, intelligent manufacturing, to realize the intelligence of various systems. To solve this traditional manual inspection of power equipment, we propose a method of intelligent identification of electrical equipment using a Single Shot Detector based on the light network—MobileNet V2. Through data enhancement and other technologies to make up for the lack of training data often encountered in deep learning. The average precision is 96.3%, and the average recall rate is 85.9% in this paper. It also has a good recognition effect for complex scenes. We provide a solution to equip unmanned aerial vehicles, remote cameras, inspection robots, and other resource-constrained devices with the power equipment identification system based on deep convolution network.