摘要
针对苹果采摘机器人视觉系统存在的问题,提出自然条件下采集到苹果图像的预处理方法,对苹果果实在图像中存在遮挡的情况做了分类,筛选和改进了所用图像处理方法,总结出了可行的快速识别和定位算法,为提高系统实时性,将该算法在嵌入式系统上实现,结果证明整个方法有效可靠。
Special for the vision system of apple harvest robot, a method was given to pretreat apple fruit images which collected in natural scenes, then screening and improving the method of the image process, after classing the occlusion which exist in the fruit images ineluctability, a algorithm for recognizing and locating the fruits was summarized, at last , aim to enhances the efficiency of the system,implementing the process in a embed system, the results shows that the process is effective and credible.
出处
《微计算机信息》
2009年第20期87-89,共3页
Control & Automation