摘要
首先把彩色图像转换至HSV彩色空间,利用图像的色相属性进行初次分割,得到初步分割结果,再将初步分割结果图像转换为灰度图像,利用该灰度图像的灰度信息进行二次分割。接下来将二次分割得到的图像转换成二值图像,再对二值图像进行腐蚀、膨胀及孔洞填充操作,分离粘连计数目标,得到适合用于计数的二值图像。最后标记二值图像的连通域,得到计数目标的个数。仿真实验结果表明,基于二次分割的彩色图像目标自动计数算法达到了精确计数的目的。
The color image is transformed into HSV color space first,and then hue is used as the key attribute to perform segmentation for the first time.Then the preliminary segmentation result is transformed into a gray image,and segmentation is performed for the second time.Next,the result of the second segmentation is transformed into a binary image.After eroding,dilating and holes filling to the binary image,connected domains are labled,and at the end came the count result of the targets.Simulation results show that the automatic counting algorithm based on two times segmentation for color targets in color image achieved the purpose of accurately counting.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第4期780-782,787,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(6087405)
关键词
彩色图像
二次分割
色相
灰度
计数
color image two times segmentation hue grayscale counting
