摘要
提出了一种基于HSV颜色空间的区域增长算法,利用统计方法得到区域增长的种子点和各点对应的分割阈值,使用网格信息寻找种子点的邻近点集,并综合考虑点云在HSV颜色空间中H分量的相近程度划分点云区域.试验表明,所提出的算法分割较精确,不容易产生过分割的情况.
A new segmentation algorithm based on HSV color space and region growing method was presented.Through dimensional histogram statistics method,the seeds and region growing threshold were extracted.Also,the grid information to find points＇ neighbors was used,and the similarity of H field of HSV space was used to determine whether the two points belong to the same region.The experiments show that the presented algorithm gains more precise result and can reduce oversegmentation phenomena.
出处
《东华大学学报（自然科学版）》
CAS
CSCD
北大核心
2014年第4期481-485,496,共6页
Journal of Donghua University(Natural Science)
基金
国家自然科学基金资助项目(81372057)
长海医院"1255"学科建设计划课题资助项目(CH125510211)
