Color correction of remote sensing images based on style transfer

Color distortion of remote sensing image is a common problem, and the color distortion of remote sensing image is not conducive to its subsequent interpretation. The existing color correction methods have problems such as poor correction effect and distortion of texture details. To solve these problems, this paper proposes an improved CAP-VSTNet style migration network. The proposed structure-color loss function can make the texture and color of the result close to the target image. The AdaIN adaptive normalization layer is introduced to improve the correction effect without increasing the calculation amount. Experiments on public datasets show that the proposed method outperforms other comparison methods in mean square error, peak signal-to-noise ratio, structural similarity and IL-NIQE(a completely blind image quality assessment method).