Abstract: In order to enhance the quality of high sampling density gray code image binarization, a threshold-belt binarization method is proposed in this study. First of all, by using this method, we need to establish a single threshold and then, according to the dispersion of the background and objectives, respectively with the single threshold, we set up a collection of threshold-belts. By removing the sampling points in the threshold belt, binary the sampling points outside of the threshold belt in order to get background and objectives, namely, the fact that the black and white stripes of the gray code image. This method avoids the phenomenon of misjudgments of the gray-scale sampling points near the single threshold. It also improves anti-interference ability of the image binarization and accuracy of the measurement.