基于形态学处理与模式识别的手部图像分割  

Hand image segmentation based on morphological processing and pattern recognition

作　　者：罗伟[1] 冯亮[1] 史桐林 张煜东[1] 

机构地区：[1]南京师范大学计算机科学与技术学院,江苏南京210023

出　　处：《现代电子技术》2015年第12期80-82,共3页Modern Electronics Technique

基　　金：江苏省三维打印装备与制造重点实验室(BM2013006);南京师范大学创新项目

摘　　要：为了更有效地分割手部图像,提出一种基于彩色阈值变换、形态学处理与模式识别的手部识别算法。利用手部与周围背景在RGB颜色分量上存在一定的差异,先通过RGB阈值变幻将彩色图像二值化,然后进行闭运算与孔洞填充,最后根据周长面积比去除背景中的孤立点集以提取手部。实验证明,该算法能够较为有效地将手部从背景中提取出来。To segment hand image effectively,a hand recognition algorithm based on color threshold transform,morphological processing and pattern recognition is proposed in this paper. Based on the differences in RGB color components between hand and its background,the color image binaryzation is performed with RGB threshold change,the close operation and hole filling are carried out,and then hand image is extracted according to the perimeterarea ratio to remove isolated point sets. The experiment results show that the proposed algorithm can extract hand from the background effectively and accurately.

关 键 词：手部图像分割 彩色阈值 形态学处理 特征提取 模式识别 

分 类 号：TN911.733.4[电子电信—通信与信息系统] TP751[电子电信—信息与通信工程]