The invention discloses an intersection traffic planning system based on a CNN detection compliance rate. The defects of violation detection in a current intelligent traffic system are analyzed, a CNN-based target detection method is adopted, the compliance rate of an intersection is calculated in real time, the intelligent police force scheduling of the intersection is achieved, and therefore, the traffic planning of the intersection is completed, the accuracy and speed of intersection violation detection are improved, the violation of vehicles and pedestrians can be detected at the same time, the intelligent scheduling of police force at the intersection is achieved by calculating the compliance rate, and manpower resources are greatly saved. According to the invention, the further development of the intelligent traffic system in combination with a current popular deep learning algorithm is facilitated, and a key technical scheme is provided for the intelligent traffic system.
本发明公开了一种基于CNN检测遵章率的路口交通规划系统,分析了当前智能交通系统中,违章检测的不足和缺陷,采用基于CNN的目标检测方法,实时计算路口的遵章率,实现路口智能警力调度,从而完成路口的交通规划,不仅提高了路口违章检测的精度和速度,而且可以同时检测车辆和行人的违章,并且通过计算遵章率实现路口警力的智能调度,极大了节省了人力资源。本专利有利于推进智能交通系统结合当前流行的深度学习算发进行进一步发展,为智能交通系统提供了关键技术方案。
Intersection traffic planning system based on CNN detection compliance rate
一种基于CNN检测遵章率的路口交通规划系统
2020-05-12
Patent
Elektronische Ressource
Chinesisch