23-06-2023 дата публикации
Номер: CN116311097A
Принадлежит:
The invention relates to the field of transportation hazardous chemical substance management and control, in particular to a hazardous chemical substance transportation vehicle abnormal behavior automatic identification and management and control technology based on deep learning and an image processing technology, and specifically relates to a hazardous chemical substance transportation vehicle abnormal behavior detection method and system, a terminal and a medium. According to the method, target detection and trajectory tracking are combined to detect the abnormal behavior of the hazardous chemical substance vehicle, so that the hazardous chemical substance vehicle can be accurately detected, the vehicle can be tracked, and the behavior of the vehicle can be judged and detected; the features of the target are fully utilized, each layer of feature map serves as output, and the problem of gradient disappearance is solved; according to the method, the target detection algorithm, the clustering ...
Подробнее