Anti-collision method and system based on laser vision fusion and deep learning

The invention provides an anti-collision method based on laser vision fusion and deep learning. The anti-collision method comprises the following steps: acquiring image information and 3D point cloud data; registering the image information and the 3D point cloud data, generating RGBD images by using the registered image data and 3D point cloud data, and dividing the RGBD images into a training set and a test set; adopting the RGBD images of the training set to train a 3D target detection model, and obtaining a trained 3D target detection model; inputting RGBD images of the test set into the trained 3D target detection model, and using a target detection method to carry out category detection on obstacles in a visual field range and calculate the distance from an object to the nearest end of a gantry crane; and according to the calculated object category and distance, determining whether deceleration and stop operations are carried out or not. According to the method, a visual means and a laser means are combi