Image Registration SIFT Algorithm Based on Adaptive Adjustment of Grayscale Weight

In view of the low color contrast of the foreground in the image of aero-engine blades, this paper optimizes the SIFT algorithm which was commonly used in image registration. The SSIM index and FSIM index in the image similarity evaluation standard are added to calculate. Using the ESFS index as the judgment standard for analysing the similarity between the color image and the corresponding grayscale image. To calculate a weight ratio which is better than the traditional algorithm. Experiments show that the proposed algorithm outperforms the traditional algorithm in saving the blade images' information, and can acquire more feature point pairs in the subsequent SIFT registration process.