摘要
采摘机器人被广泛用于各种农产品的收获,能显著降低农业生产中的人力成本,解决目前农业生产中劳动力匮乏的问题。计算机视觉技术可以将成熟的瓜果从其自然生长环境中识别出来,是采摘机器人实现目标定位和协调机械手的关键。针对目前采摘机器人受瓜果不同形态特征、质地影响和通用性较低等问题,通过优化公式算法和模型,设计出了基于计算机视觉的采摘系统,并对黄瓜、西红柿和柑桔这3种瓜果进行图像识别和采摘。试验结果发现:不同瓜果图像整体识别率达到80%以上,对图像中重叠遮挡的瓜果个体识别率也较高,均超过90%。该计算机视觉系统装载到采摘机器人上后,在自然条件下对3种瓜果采摘成功率在70%以上,可实现对不同瓜果的高效采摘,为采摘机器人更好地满足农业生产的实际需求提供了研究基础。
Picking robots are widely used in the harvest of various agricultural products, can significantly reduce the labor costs of agricultural production, agricultural production to solve the current problem of lack of labor. Computer vision technology can be mature fruit from its natural growth environment to identify out, picking robots to achieve the goal of positioning and coordination of the robot key. In order to solve the problem that the picking robot has different morphologi- cal characteristics, texture effects and low versatility, this paper designs a computer-vision-based picking system by opti- mizing the formula algorithm and model, and cucumber, tomato and citrus. The results showed that the overall recognition rate of different fruits was more than 80%, and the individual recognition rate of overlapping fruit was higher than 90%. When the computer vision system is loaded on the picking robot, the success rate of picking three kinds of fruits under natural conditions is more than 70% , which can realize the efficient picking of different fruits, and better meet the actual demand of agricultural production for the picking robot providing the basis for research.
出处
《农机化研究》
北大核心
2018年第1期200-203,共4页
Journal of Agricultural Mechanization Research
基金
河北省科学技术厅项目(20150358)
作者简介
王彦辉（1980-），男，河北高阳人，讲师，硕士，（E—mail）wangyanhui8524@163．com。
