一种用于尿沉渣图像的自适应阈值分割新方法  

An Adaptive Threshloding Segmentation Method for Urinary Sediment Image

作　　者：李勇明[1] 曾孝平[1] 覃剑[1] 韩亮[1] 

机构地区：[1]重庆大学通信工程学院,重庆400030

出　　处：《生物医学工程学杂志》2009年第1期6-9,共4页Journal of Biomedical Engineering

基　　金：国家自然科学基金资助项目(30570473)

摘　　要：我们针对复杂散焦的尿沉渣图像的精细分割,首先使用小波变换和形态学处理消除散焦影响并进行图像的粗分割,然后对粗分割得到的小波变换子图像进行自适应阈值处理,结合形态学处理完成细分割,最后再采用剥离算法处理粘连重叠成分。该方法不受散焦影响,充分利用了图像的多种信息,实验结果表明,该方法对尿沉渣图像的分割有效且令人满意。In this paper is proposed a new method to solve the segmentation of the complicated defocusing urinary sediment image. The main points of the method are： （1） using wavelet transforms and morphology to erase the effect of defocusing and realize the first segmentation, （2） using adaptive threshold processing in accordance to the subimages after wavelet processing, and （3） using ＇peel off＇ algorithm to deal with the overlapped cells＇ segmenta- tions. The experimental results showed that this method was not affected by the defocusing, and it made good use of many kinds of characteristics of the images. So this new mehtod can get very precise segmentation; it is effective for defocusing urinary sediment image segmentation.

关 键 词：尿沉渣 图像分割 小波变换 数学形态学 自适应阈值 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]