摘要
研究了灰度图象的二维ｏｔｓｕ自动阀值分割法．根据灰度图象的象索点灰度，邻域均值二维直方图的分布特点，给出了一个新的分割阀值判决函数．理论分析与实验结果表明：该阀值判决函数比仅使用一个二维矢量阀值具有更强的抗噪声能力，分割性能更好．
Two dimensional otsu thresholding for grey image segmentation was discussed. A new thresholding output function was discussed. A new thresholding output function was given based on the distribution features of 2 D histogram (grey level/local average level). The Theoritical analysis and comparative experimental results show that this new function performs are much better when the image are corrupted by noise.
出处
《西南师范大学学报（自然科学版）》
CAS
CSCD
北大核心
1998年第6期658-662,共5页
Journal of Southwest China Normal University(Natural Science Edition)
关键词
灰度图象
直方图
阀值
OTSU法
图象分割
grey image
histogram
threshold value
otsu threshold
image segmentation
