开放式茄子采摘机器人设计与试验

Design and Experiment of Opening Picking Robot for Eggplant

根据茄子生长的空间分布,利用优化设计方法进行了机器人本体结构参数设计,开发了4自由度关节式采摘机器人机械本体。采用基于直方图的固定双阈值法对G-B灰度图像进行分割。根据果蔬采摘机器人对视觉系统的要求,提取了果实目标的轮廓、面积、质心、外接矩形以及切断点等特征。整机性能测试结果表明：以单摄像头两步法测距,当测量距离在275~575 mm范围内,测量误差基本都在±18 mm之内;整机试验系统运行稳定可靠,抓取成功率为89%,平均耗时37.4 s。

The optimal design was made on the parameters of the robot framework and structure according to the growth distribution space of eggplant. A joint type robot with 4-degree freedom for fruit and vegetable picking was developed. The fixed bilateral threshold based histogram was adopted to split the G-B gray images. To meet the vision requirement of the fruit and vegetable picking robot, the object＇s characters, such as outline, area, center of gravity, enclosing rectangle and the point of cutting off, were picked up. The performance test on the whole machine was done on the robot system with the method of single camera and two-step measuring distance, which shows that the measure error is within ± 18 mm when the measure distance ranges from 275 mm to 575 mm. The system runs smoothly and reliably. The success ratio of picking is 89 % and the average time consumed is 37.4 s.