Abstract: In the high-precision measurement system, the CCD pixes nonuniformity will give a greater margin of error to measurement results. According to the characteristics of noise leading to the image pixels nonuniformity, this article establishes the corresponding image model. Adaptive threshold segmentation algorithm is used for image binarization and isolating the effective pixels. An algorithm using two points linear method for pixel non-uniformity correction is put forward. Experiments adopt the high speed area array CCD to gather images with the circular facula launched by the laser, and verify the algorithm by calculating the facula center coordinates of the image. The results showed that the algorithm can separate light intensity positions from the image, correct pixel nonuniformity, and stabilize pixel gray value of light intensity positions. Under the same conditions for image acquisition, the center coordinates of image light intensity positions stabilize.