动态有序变化的阈值曲面方法

对于背景灰度不一致的图像的二值化问题,可采用在图像的不同区域选取不同阈值的动态阈值曲面方法解决。本文提出一种有序变化的动态阈值曲面方法,通过选取三个动态变化调节系数,使曲面动态地逼近阈值分割的最佳值,同时又降低曲面的误分割,能较好地提取二值图像。

To thresholding an image with uneven background, different thresholding values should be selected for different zones in the image. We propose an active thresholding method, which push the thresholding surface to optimal position and decrease some mis-thresholding by choosing three adjustment coefficients. Experiment results show that the proposed method can segment image well.