基于小波变换和拉普拉斯算子的细胞图像边缘检测方法  

Edge Detection of Blood Cell Image Based on Wavelet Transform and Laplace Operator

作　　者：陆琳[1] 孙福军[1] 周小平[1] 刘国传[1] 

机构地区：[1]中华人民共和国北京出入境检验检疫局,北京100026

出　　处：《中国医学物理学杂志》2007年第4期288-289,299,共3页Chinese Journal of Medical Physics

基　　金：国家自然科学基金重点项目(39889001)资助

摘　　要：目的:介绍一种基于小波变换和拉普拉斯算子的血液细胞图像边缘识别方法。材料和方法:取正常人血样5mL,制成血液细胞图片12片,对血图像进行预处理后,利用小波变换的多分辨率特性滤除细胞图像中的干扰成份,根据血液细胞边缘附近的灰度分布梯度较大的特点,采用拉普拉斯算子及双阈值法对其进行边缘检测和识别。结果和结论:实验结果表明,结合小波变换和拉普拉斯算子的边缘提取算法对血液细胞图像边缘提取有良好的效果,为下一步对血液细胞的形态学分析、分类和识别提供了新途径。Objective： This paper introduces an edge detection method of blood cell image, which is based on （wavelet transform WT） and （ laplace operator LO）. Materials and Methods：Using about 5mL blood samples of healthy people, the author made 12 blood smears （each of them contains 100 blood cells approximately）. First, the method utilizes the multi-resolution characteristic of WT to remove main interferences in blood cell images,and then extracts blood cells＇edges by LO and bi-thresholds method,according to the blood cells＇features of grey level distribution. Results and Conclusion：Our experimental results demonstrate that this algorithm is an efficient method to detect the edges of blood cells and a new approach for further analysis of blood cells can be proposed.

关 键 词：小波变换 拉普拉斯算子 血液细胞 边缘检测 

分 类 号：R319[医药卫生—基础医学]