Environment perception method based on deep learning

• Main Authors: WU AOYUN, LI HUI, ZHANG XIUMEI, MA YUMIN, LIU YUE, LIU FANG, YANG FAN, LUO MINGYUE
• Format: Patent
• Language: Chinese; English
• Published: 19.09.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses an environment perception method based on deep learning, and relates to the field of image data processing, deep learning and detection identification. The method comprises the following environmental perception detection steps: establishing an elevator target special data set; elevator state images are marked; a YOLO V5 network structure is improved; training an elevator door state target special data set; and the improved YOLO V5 network model is used for carrying out environment perception detection on the elevator door target. The method can be applied to the field of image data processing, and compared with the prior art, the training amount of the model is reduced, resources occupied by the model are reduced, meanwhile, high precision and speed are kept, and the environment perception requirement of a mobile robot platform for the elevator door state model is met. 本发明公开了一种基于深度学习的环境感知方法，涉及图像数据处理、深度学习和检测识别领域。其中包括以下环境感知检测步骤：建立电梯目标专有数据集；标注电梯状态图像；改进YOLO V5网络结构；训练电梯门状态目标专有数据集；利用改进的YOLO