基于视觉识别技术的移动式樱桃采摘机器人设计

为了降低樱桃采摘过程中的破碎率,提高采摘效率,设计一种新的基于视觉识别技术的移动式樱桃采摘机器人。将TI公司的32位数字处理器芯片TMS320F2812作为核心处理器设计移动式樱桃采摘机器人,本文给出硬件总体方案,并且通过SAA7105完成对视频图像的编码操作。软件设计主要包括樱桃的定位、机器人手眼标定和机器人路线规划。依据目标上多个点的三维位置信息对樱桃球模型进行重建,通过最小二乘法求出樱桃质心位置,实现樱桃的定位。将樱桃在机器人坐标系中的定位简化成求解摄像机坐标系和机器人坐标之间的映射矩阵,通过映射矩阵实现机器人手眼标定。考虑到樱桃的采摘要求,通过分段路径规划对机器人的运动进行控制。实验结果表明,所设计机器人图像处理性能强、定位精度高,在保证采摘精度的同时,有很高的采摘效率。

Mobile cherry picking robot based on visual identity design

in order to reduce the breakage rate in the cherry picking process,improve the efficiency of picking,design a new mobile cherry picking robot based on visual recognition technology,the TI company"s 32-bit digital processor chip TMS320F2812 as the core processor design mobile cherry picking robot,gives the hardware overall scheme,complete the video image through SAA7105 coding operation.Software design mainly includes the cherry and robot hand-eye calibration and robot path planning.According to target multiple points on the three dimensional position information of cherry ball model reconstruction,calculated by the least squares method the barycentric position of the cherry,cherry"s positioning.To simplify the cherry in the robot coordinate positioning into solving the camera coordinate system and robot coordinate mapping between the matrix and robot hand-eye calibration was achieved by mapping matrix.Considering the cherry picking requirements,through the section to control the movement of the robot path planning.The experimental results show that the designed robot image processing performance is strong,high positioning accuracy,while picking precision,a high degree of picking efficiency.