一种改进的基于大津自动阈值选择法的分窗二值化过程

A Modified Process of Binarization in Partitioned windows Based on Otsu's Automatic Threshold Selection Method

利用大津自动阀值选择法进行分窗二值化.是人们在对图象进行二值化处理肘为了克服光照不均等因素的干扰而广泛采用的方法,但是,分窗的大小往往会带来窗口内仅含背景(或对象物)的问题,严重影响二值化的正确性.本文给出了一种基于差分和的金字塔数据结构判别过程,有效地解决了这一问题,进而能得到令人满意的二值化效果.

Otsu's automatic threshold selectionmethod is widely used in eleminating thedisturbance of the uneven lighting duringthe binarization in partitioned windows.Butthe size of the window may often bringabout the problem of homogenous windows,i.e.,windows contain only background orobject pixels,and this can seriously affectsthe result of the binarization.In thispaper,we put forward a discriminatingmethod which is based on pyramid datastructure of the ∑-grad values,and byusing the method the droblem is solvedperfectly.