Intelligent interactive test system for the human-machine interface function of electric vehicle charging piles based on computer vision

In recent years, the market share of new energy vehicles has been increasing, so the number of electric vehicles charging piles put into use has been rising. Therefore, the inspection work of the charging pile before leaving the factory is very important. The human-machine interface function is a necessary test item, the current testing method of which is still mainly manual, with low efficiency, high error rate, high cost, and the problem of easy electric shock. In order to solve these problems, this paper proposes an intelligent interactive test system for the human-machine interface function of electric vehicle charging piles based on computer vision to replace manual labor. The system consists of three parts: the interface perception module, the robotic arm control module and the cloud platform. This paper also puts forward the key technologies in the interface perception module, including the multi-task network that completes the interface positioning and text detection, and the parallel multi-model text recognition network. The system has been verified by twenty devices in the quality inspection workshop of the State Grid of China, which can well complete the quality inspection of charging piles. The paper also provides the comparison results with other quality inspection algorithms.