基于改进型色度学算法的细胞分割研究与分析  

Segmentation and Analysis of Cells Based on Improved Colorimetry Algorithms

作　　者：刘玉成[1] 杨露鑫 张颖超[3] 

机构地区：[1]南京财经大学国家级实验教学研究中心,江苏南京210023 [2]南京财经大学,江苏南京210023 [3]南京信息工程大学,江苏南京210044

出　　处：《凯里学院学报》2017年第6期81-86,共6页Journal of Kaili University

基　　金：江苏省六大人才高峰(2106-A-027);南京财经大学高教所研究课题(2016-GJ-033)

摘　　要：由于尿沉渣中细胞图像的自身特点和采集过程中受光源等环境的影响,细胞彩色图像的采集经常会发生彩色偏移的问题,再加上图像背景本身的复杂性,使得对尿沉渣中有形成分的分割和处理非常困难.为改变传统细胞图像分割和处理方法的不足,采用了一种改进型的色度学的图像处理方法.系统采用白平衡算法,针对因外部光线所造成的误差进行校正,从而解决了由光照引起的图像彩色偏移问题.接着利用改进型RGB向量空间分割法对图像进行处理,在复杂的背景中将尿沉渣有形成分分割出来.最后得到了理想的实验结果,进而使得分割和处理的效果更理想,性能更稳定.Due to the features of the cell image in the urinary sediment and such environmental influence as light sources on the collection process, the collection of color images of cells often experiences color offsets, which, coupled with the complexity of image background, made the segmentation and treatment of the formed element in the urinary sediment extremely hard. In order to remedy the inadequacies in the traditional segmentation and treatment of cell image, the paper adopted an improved colorimetric image treatment method, this system adopted a White Balance method to adjust the discrepancy caused by external light and solves the image color offsets due to exposure to light, then process the image with a RGB vector space segmentation method, and segment the formed element from the urinary sediment against a complex background. The method was able to achieve a more ideal segmentation and treatment and a more stable performance.

关 键 词：向量空间法 MAHALANOBIS距离 特征提取 色度学 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]