The invention relates to a
video image sharpness
processing method in a
fog and
haze day and a device thereof. The method comprises the following steps: establishing a
quantitative model between an atmospheric transmission parameter and an
optical propagation medium parameter; obtaining the
optical propagation medium parameter of a practical
measurement site, and calculating an atmospheric quantitative transmission parameter by utilizing the
quantitative model; carrying out downsampling
processing to an original image which contains
fog and
haze to obtain a preprocessing template image, and extracting a dark-channel image; extracting the atmospheric optical value A of the original image I(x) which contains the
fog and
haze, and initializing a depth-of-field correction factor (
Omega); optimizing A and
Omega in a transmissivity graph of the I(x); estimating a template transmissivity graph t<model> of a preprocessing image, and carrying out
upsampling to obtain a transmissivity graph t(x); and according to the t(x) and the I(x), calculating a sharpness
processing result image. A fog and haze image sharpness parameter is subjected to
closed loop correction through the
quantitative model, an image sampling and multi-scale interpolation method is combined to avoid defects that parameters can not be accurately artificially defined and can not be regulated according to practical environment in a dark channel image defogging process, and a sharpness processing effect is improved.