Spectral depth perception evaluation training method based on image-background separation

The invention provides a spectral depth perception evaluation training method based on image-background separation, and relates to the technical field of depth perception evaluation, under the cooperation of background canvas, foreground image models correspond to different spectral depth concave-convex relationships, and an observer observes and feeds back the foreground image models through observation equipment, so that the spectral depth perception evaluation accuracy is improved. The observed concave-convex relationship is the spectral depth concave-convex relationship corresponding to the foreground image model which is perceived by self depth perception, and comparing the concave-convex relationship of the foreground image model which is observed and fed back by an observer through observation equipment with the preset spectral depth concave-convex relationship corresponding to the foreground image model; according to the scheme, the observer is helped to find the abnormal condition of the spectral dep