Foggy-day image sharpening method of multi-scale Retinex model based on HIS space

The invention relates to a foggy-day image sharpening method of a multi-scale Retinex model based on an HIS space. The foggy-day image sharpening method comprises the six main steps of: (1) acquiring a source image, if the source image is a black-white image, converting the source image into double type from byte type; if the source image is a color image, converting the source image into double type from byte type respectively by black-white pixel values of R (red), G (green) and B (blue) channels; (2) converting the R, G and B channels to H, I and S; (3) respectively carrying out improved multi-scale retinex (MSR) algorithm on the pixel points obtained by the step (2), and obtaining a new image; (4) carrying out contrast spreading on the image obtained by the step (3); (5) respectively converting H, I and S obtained by the step (4) to R, G and B; and (6) carrying out synthesis on R, G and B obtained by the step (5), and displaying the sharpened image. In the foggy-day image sharpening method, the sharpening