Research on the Environment Perception System of Intelligent Connected Vehicles Based on Machine Vision

In this paper, a Kalman filter is proposed to detect the environment of intelligent connected vehicles. The system improves the accuracy and robustness of environment perception by fusing information from multiple sensors (such as cameras, lidar, etc.). In the system design, the Kalman filter is used to track dynamic targets on the road in real time, such as pedestrians, other vehicles, and obstacles, and ensures accurate estimation of the target position through prediction and update steps. In addition, the fusion of Kalman filter and other sensor data further enhances the accuracy of environmental modeling. Experimental results show that the proposed system performs well in multiple typical environments, especially in complex urban roads and severe weather conditions. The Kalman filter algorithm can effectively suppress noise and maintain high tracking accuracy. The object detection accuracy on the KITTI dataset reaches 96.2%, with good real-time performance and a delay of only 40ms. The experiment verifies the effectiveness of the environment perception system based on the Kalman filter, providing strong technical support for the safety and reliability of intelligent connected vehicles.