High-quality image synthesis method based on diffusion model

The invention provides a high-quality image synthesis method based on a diffusion model, and the method comprises the steps: carrying out the fusion and conversion of a plurality of foreground and background images into a synthetic image with a high sense of reality through a gradual addition and noise removal mode of the diffusion model; according to the method, an edge contour correction model is introduced, the edge contour information of a foreground is used as a prompt, the diffusion process of the edge contour information is stabilized, and it is guaranteed that the edge of a synthesized image is excessively smooth under the condition that overlapped objects exist; according to the method, a color conversion model based on a ControlNet is further introduced, color information of a foreground and a background is processed and then fed back to the ControlNet to serve as a prompt of a diffusion model decoder, and it is guaranteed that the color of a synthesized image is closer to the reality under the cond