浙江农业学报

• 生物系统工程 • 上一篇    下一篇

基于RGB颜色相似度的成熟草莓图像分割

  

  1. (1. 武汉轻工大学 数学与计算机学院,湖北 武汉 430048;2. 华中科技大学 自动化学院,湖北 武汉 430074;3. 南昌航空大学 测试与光电工程学院,江西 南昌 330063)
  • 出版日期:2016-02-25 发布日期:2016-03-10

Image segmentation of ripe strawberry based on RGB color similarity

  1. (1. School of Mathematics and Computer, Wuhan Polytechnic University, Wuhan 430048, China; 2. School of Automation, Huazhong University of Science and Technology, Wuhan 430074, China; 3. School of Measuring and Optical Engineering, Nanchang Hangkong University, Nanchang 330063, China)
  • Online:2016-02-25 Published:2016-03-10

摘要: 针对复杂背景下的成熟草莓图像,提出了一种基于RGB颜色相似度的成熟草莓图像分割算法(CS\|BASED RSIS)。首先提取成熟草莓区域,确定成熟草莓的主颜色,然后遍历待分割的图像,求出每个像素点相对于主颜色的颜色比和相似度,进行颜色相似度的阈值分类,最后经多次膨胀和去除小面积对象的数学形态学处理,输出分割结果。试验结果表明,在无粘连无遮挡、无粘连有遮挡、有粘连有遮挡3种复杂环境下,与OTSU等图像分割算法相比,CS\|BASED RSIS算法不仅能达到更好的分割效果,而且平均分割时间仅为0965 s,能满足成熟草莓机械化采摘的实时性要求。

关键词: 图像分割, 颜色相似度, 成熟草莓

Abstract: In view of the limitations of traditional simple threshold algorithms under the complicated background, a ripe strawberry image segmentation algorithm based on RGB color similarity (CS\|BASED RSIS) was proposed. Firstly, ripe strawberry regions were extracted, which determined the main color of ripe strawberry. Then, the CS\|BASED RSIS algorithm traveled all pixels in the image for segmentation, calculated the color ratio and similarity of each pixel point relative to the main color, and threshold classification on color similarity were carried out. Finally, the image was processed with the multiple expansion and removing small objects and the segmentation result was output. Experimental results showed that, in three kinds of complex environment, i.e., no adhesion and no occlusion, no adhesion and occlusion, adhesion and occlusion, the CS\|BASED RSIS algorithm not only could achieve better segmentation effect compared with the OTSU image segmentation algorithm, etc, but also could save time, its average segmentation time was only 0965 s, and the CS\|BASED RSIS algorithm could meet the real\|time requirement of ripe strawberry mechanization picking.

Key words: image segmentation, color similarity, ripe strawberry