摘要
通过对青椒RGB模型、颜色因子、直方图阈值的分析,提出了一种颜色特征与直方图阈值相结合进行田间青椒图像分割的方法,该方法无需灰度转换。试验结果表明,该算法能很好地从图像背景中分离田间青椒果实,并较好地保存青椒轮廓信息,分割成功率高于85%。进一步对分割的图像进行平滑,形态学处理,如膨胀、腐蚀等,可以有效消除孔洞现象,有利于对青椒的进一步识别。
In this paper, by means of the analysis of the green pepper RGB model, color factor and histogram threshold value, a method that combines the color characteristic with the histogram threshold value is introduced to cut up the green pepper image in the field. This method does not need gray-scale conversion. The result of the experiment indicates that the field green pepper fruit can be separated well from the picture background by this algorithm, and the outline information of the green pepper is preserved well and the division success effect is greater than 85%. Furthermore, through the smooth and morphology processing to the segmentation image, such as the inflation, the corrosion and so on. For the division image, the holes phenomenon can be effectively eliminated, and the green pepper can be future identified.
出处
《微型机与应用》
2010年第4期51-53,共3页
Microcomputer & Its Applications
关键词
颜色特征
图像分割
采摘机器人
阈值
color characteristics
image segmentation
picking the robot
threshold value
作者简介
于杨，女，1980年生，硕士研究生，主要研究方向：计算机控制
崔天时，男，1967年生，博士，副教授，主要研究方向：信息处理与智能测控
董桂菊，女，1967年生，硕士，副教授，主要研究方向：信息处理与智能测控。
