The bionic compound eye system solves the contradiction between the field of view and resolution in the traditional single-aperture optical system, and has both large field of view and high resolution. However, with the increase of field of view and resolution, the image information increases, which brings the problem of efficiency and quality in image Mosaic. To solve this problem, a bionic compound eye multi-channel image Mosaic fusion method based on improved scale-invariant feature transformation (SIFT) algorithm and principal component analysis (PCA) algorithm is proposed. The method reduces the feature point extraction area, reduces the number of multiplexed image feature point matching, and decreases the dimensionality of image feature point descriptors. The improved adaptive iterative random sampling consistent (RANSAC) algorithm is used to purify the feature points to increase the robustness, and finally the high-quality fusion of multiple sub-images by the weighted average algorithm is completed. The experimental results show that the proposed algorithm is reasonable in design, and with the increase of the complexity of image information, the Mosaic efficiency is improved compared with the traditional algorithm, and at the same time, the Mosaic quality is better. It can effectively Mosaic and integrate the multi-channel images of the bionic compound eye system, and provides a reference for the mosaic and fusion of the multi-channel image system.