With the continuous evolution of intelligent detection algorithm based on deep learning method, its network structure is constantly evolving, and the degree of practicability is constantly improved. Therefore, the feasibility of applying it to complex battlefield environment to form practical intelligent perception ability is constantly improved. However, the reliability and interpretability of the algorithm have not been solved. In the future, in the framework of the ground unmanned platform system, the target detection and recognition method based on deep learning is used to fuse the sensing signals of multiple sensors to explore how to reliably collect the conditions of enemy vehicles, personnel, related objects near the unmanned platform, as well as the geographical and meteorological environment in the visual range. So as to realize the multiple intelligent sensing process and realize the perception and understanding of complex battlefield environment for unmanned platform. It will be the main task of our army's intelligent perception theory in the future to provide technical reserves for autonomous environment determination, autonomous walking, autonomous danger determination and even threat automatic disposal.
李 程,夏 丹*,董世运,胡雪松,戴 迪.【专题】复杂陆战场环境下的智能感知理论现状与发展[J].国防科技,2021,42(3):42-48；LI Cheng, XIA Dan, DONG Shiyun, HU Xuesong, DAI Di. Current situation and future development of intelligent perception theory in complex land battlefield environment[J]. National Defense Technology,2021,42(3):42-48复制