自动提取植物染色体特征的研究

Research on Automatic Feature Extraction of Plant Chromosome

植物染色体特征参量的提取是实现植物染色体计算机自动识别和分析的关键,本文首先对原始植物染色体图像进行图像分割、边界搜索、背景噪声滤除等处理,将各条染色体从图像中提取出来。然后,提出采用边界的高阶方向链码来判定染色体边界上明显凹凸点的分布,又根据染色体的细化中轴,自动判定出染色体着丝点的位置。进而,测量得到描述染色体特征的主要参量。经对植物染色体图像进行实验,其结果表明这一系列处理方法对植物染色体特征的自动提取是有效的。

Automatic feature extraction of plant chromosome is the kcy for realizing automatic plant chromosome analysis and recognition. This paper first preprocess the primitive plant chromosome image. The preprocessing include image segmentation, boundary tracing, background noisy filtering, etc. Then using high order chain code of boundary to detect the convex and concave points of the boundary. After that, automatically detect the location of chromosome centromere according to the thinning medial axis. At last, compute all the parameters of chromosome features. The result was satisfying by experimenting on plant chromosome image and this shows that all the series of processing methods are basically effective for automatic feature extraction.