基于Matlab的根系断层图像分割处理算法  

Study on Matlab-based Segmentation Algorithms for X-ray CT Image of Plant Root

作　　者：周志艳[1] 周学成[1] 罗锡文[1] 李庆[1] 

机构地区：[1]华南农业大学工程学院

出　　处：《农业机械学报》2006年第12期118-121,共4页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金资助项目(项目编号:60375005)

摘　　要：根据植物根系原位形态可视化研究的需要,利用Matlab提供的图像处理算法,编制了根系断层图像分割处理算法的优选整合程序,用多种不同算法组合对根系CT原位图像进行实例检验,采用最终测量精度法作为评价准则,对分割效果进行分析研究。研究结果表明,在90%的置信水平下,处理模块的不同组合顺序以及每个模块中不同算法的选择对最终分割效果有显著的影响,各因素对分割效果总变异贡献率的大小排序为分割(46.1%)、灰度调整(17.9%)、分割前滤波(4.5%)、分割后滤波(2.6%)、执行序列(0.5%)。根据所设计的实验例证,提出了处理流程及算法中的较优组合。According to the need of plant root morphology visualization, an optimal selection system for segmentation algorithms of plant root X-ray CT image based on Matlab has been developed, and the segmentation experiments with real image instance were provided to show the effectiveness of the system. The analysis based on the evaluation of algorithm （UMA ultimate measurement accuracy） was developed. At the 90% confidence level, the results showed that the effect of every factor to the final segmentation quality was significant. The order of the factors＇ percent contribution was. segmentation （46.1%）, intensity adjustment （17.9%）, and noise reduction filtering before segmentation （4.5%）, noise reduction filtering after segmentation （2.6%） and treatment combinations in different order （0. 5 %）. This system could be used as an examination to find out the best treatment algorithm combination under the given experimental instances.

关 键 词：根系 图像分割 算法优选 

分 类 号：S126[农业科学—农业基础科学] TP391.41[自动化与计算机技术—计算机应用技术]