Point cloud dynamic object filtering method based on camera and laser radar fusion

The invention relates to a point cloud dynamic object filtering method based on camera and laser radar fusion, the method is based on a data acquisition platform, the data acquisition platform is provided with a camera and a laser radar, and the method comprises the following steps: obtaining a real-time point cloud through the laser radar, and synchronously obtaining a real-time image through the camera, the camera and the laser radar adopt a target-free external parameter automatic calibration method to pre-calibrate external parameters, a dynamic object detection frame in an image is obtained through a visual target detection network based on deep learning, point cloud is projected into the image according to the external parameters obtained through pre-calibration, and a dynamic object detection frame is obtained by combining the point cloud projected onto the image. And after the point clouds in the dynamic object detection frame are screened out, clustering segmentation is carried out on the point cloud