An improved Otsu method for threshold segmentation based on set mapping and trapezoid region intercept histogram

Abstract

The line intercept histogram Otsu method enjoys a good segmentation performance but relatively poor computational efficiency and anti-noise capability. This paper analyzes its set mapping essence, and then proposes a new Otsu method for threshold segmentation based on set mapping and trapezoid region intercept histogram. Firstly, in order to improve its computational efficiency, a mapping rule for the mapping of the pixels in the two-dimensional (2D) histogram onto different trapezoid regions is established to compress its threshold space sufficiently. Then the post-processing strategy is introduced to modify the mapping rule, so that the algorithm can better maintain the image details and resist noises. The experimental results show that the algorithm has a better performance in computational efficiency, and can achieve a balance between the noise elimination and the detail retention in image segmentation.