Multi-channel video image splicing method

The invention discloses a multi-channel video image splicing method, which comprises the following steps of: acquiring image data of each channel of video at the same moment, splicing images, and calculating a pixel coordinate position mapping transformation matrix and a pixel value transformation mask matrix from an original image to a spliced image; acquiring images in a video stream during actual splicing, carrying out mapping transformation and mask matrix transformation on each frame of image by using a kernel function, realizing mapping transformation on pixel coordinate positions, and realizing mask matrix transformation on pixel values, so as to obtain a spliced image; and calling the trained model to carry out target identification and tracking on the spliced image. The splicing principle is fully understood, the splicing process is converted into transformation of pixel point positions and pixel values of all paths of images from a mathematical angle, a mapping matrix of pixel position coordinates a