The star map has the characteristics of large amount of data and wide angle of view. The star points are displayed as limited pixels on the image and easy to get disturbed by noise, so it is difficult to build a feature model to obtain the complete star map. In this paper, we have revised the OTSU algorithm and extracted the star points accurately according to the characteristics of star point pixels. We are inspired by the idea of compressed perception to break the local spatial relationship of stars, construct the aggregated star group model of relative relationship of local stars, realize the matching and construct the extended model to realize the stitching of wide angle star images. In order to meet the requirements of efficient transmission and storage in the air or on the ground, we have proposed the algorithm of star point storage and mapping. We set up a database with 600 frames of real star images from CCD camera and 600 frames of simulated star images, and the experiments show that our algorithm can compress the data and realize the star image stitching.