基于粒度测量的重叠圆形颗粒图像分离方法

为了解决颗粒图像（例如红细胞图像）处理中经常出现的重叠现象，即多个颗粒重叠或者黏连在一起，提出一种分离重叠的圆形颗粒图像的方法.采用粒度测量获取图像中圆形颗粒的半径信息，计算重叠区域外轮廓的链码序列， 然后利用半径和轮廓链码来估计各个单独颗粒的中心区域以及重叠区域中颗粒的数目，接着从各个颗粒中心区域出发逐渐膨胀，在对应不同颗粒的区域开始接触的位置设置分离点，从而达到分离重叠颗粒的目的.以重叠红细胞图像为实验对象，应用该方法进行图像分离，并和分水岭算法的分割结果进行比较，实验结果表明，该方法是有效可行的.

Separating algorithm for overlapping circular granule images based on granulometry

A separating algorithm was proposed to resolve the problem of overlapping phenomena on appearing in granule images such as red blood cell. Granulometry was used to estimate the circular granule's radius. The chain code sequences of the contours of overlapping regions were calculated. With radius and contour chain codes, the central regions of every single circular granule and the number of granule were determined. Then, morphological dilation was continuously applied on every single granule's central region, during which, every region corresponding to different granule expanded gradually. The points where different regions begin to join were mapped as separating points. As an example, the algorithm was used to separate overlapping cells in red blood images. Its result was compared with the result of the algorithm based on watershed. Experimental results show that the proposed algorithm is relatively more effective than other algorithms.