bao zhongxu / China University of Mining and Technology
牛 强 / 中国矿业大学计算机学院
Edge visual perception technology is a research hotspot in the field of coal mine safety monitoring, but the complex environment of high temperature, high humidity and high dust in coal mines makes it difficult for the existing edge visual perception technology to be directly applied to underground coal mines, mainly facing the following problems: first, the network and computing resources of edge computing equipment in underground coal mines are limited, which makes it difficult to balance the accuracy of underground edge visual perception models and the hardware constraints of equipment; Second, the high temperature, high humidity and intrinsic safety requirements of the coal mine are subject to high environmental constraints, resulting in the high temperature and unstable operation of edge computing equipment. In order to solve the above problems, this paper combines neural architecture search and selective reasoning methods to study the equilibrium method of visual perception computing at the edge of the mine