Expressway vehicle illegal parking detection method based on semantic segmentation and target detection

The invention discloses an expressway vehicle illegal parking detection method based on semantic segmentation and target detection, and the method comprises the steps: obtaining a road background image of an offline monitoring video through Gaussian background modeling, dividing a road region through a semantic segmentation neural network, and eliminating the subsequent non-expressway false detection. Afterwards, for a traffic video to be detected under the same camera, the motion state of the vehicle is obtained through target detection and tracking, and the illegal parking vehicle is detected through judgment of the motion state of the vehicle. The invention designs a detection method capable of automatically judging vehicle illegal parking events, and an excellent vehicle illegal parking detection effect is achieved in a real highway traffic video. 本发明公开了一种基于语义分割和目标检测的高速公路车辆违停检测方法，首先，利用高斯背景建模获取离线监控视频的道路背景图，再使用语义分割神经网络划分出道路区域，排除后续非高速公路的误检。之后针对同一摄像头下要检测的交通视频，利用目标检测和跟踪获取车辆的运动状态，通过车辆运动状态的判别从而检测出违停车辆。本发明设计了一种可以