视觉传感器在采摘机器人目标果实识别系统中的应用

The Application of Visual Sensor in the Recognition System of the Target Fruit of the Picking Robot

目标果实的识别和定位是准确采摘的核心,是采摘作业的必要条件。为此,首先从硬件和软件两方面对采摘机器人总体框架进行了介绍,然后搭建了双目视觉系统,并建立了摄像机标定模型,最后实现了对目标果实的识别和定位功能。试验结果表明:采摘机器人目标果实识别系统对目标果实的识别定位误差在8mm以内,成功采摘率在96%以上,系统精度高,采摘效果良好,能够满足采摘机器人的作业要求,对实现水果采摘的自动化、无人化具有重要现实意义。

The recognition and location of the target fruit is the core of the accurate picking, and it is the necessary condition for the picking operation. First of all, the general framework of picking robot is introduced from two sides of hardware and software. Then is established binocular vision system and camera calibration model. Finally, it realized the recognition and location function of target fruit. The experimental results show that the recognition error of picking robot target recognition system of fruit picking within 8mm, success rate is more than 96%, the system has high precision, good picking effect, which can meet the operational requirements of the picking robot. It has important practical significance to realize the automatic unmanned fruit picking.