非结构环境中扰动葡萄采摘点的视觉定位技术

发布时间：2018-02-14 12:16

本文关键词： 葡萄扰动采摘点视觉定位　出处：《农业机械学报》2017年04期 　论文类型：期刊论文

【摘要】：非结构环境下采摘机器人对扰动葡萄采摘点的视觉识别定位有实际应用价值。首先基于"挠性杆-铰链-刚性杆-质量球"模型对葡萄串形态进行分析,将其扰动状态分解为XOY和YOZ 2个平面的类单摆运动,对扰动葡萄进行视频拍摄,通过对视频中多帧葡萄图像进行Otsu阈值分割得到果实和果梗,并计算出各帧图像中葡萄串的质心;对各帧图像的葡萄串质心进行曲线拟合,计算出葡萄类单摆运动的周期与摆角,从而确定当前扰动葡萄是否适合视觉定位;对可实现视觉定位的扰动葡萄,选取类单摆中间位置质心点对应的葡萄图像,对葡萄串上方矩形区域进行Canny边缘检测,再利用霍夫直线拟合结合角度约束法实现扰动葡萄采摘点的定位。视觉定位试验结果表明:自然环境中不同光照下扰动葡萄采摘点的视觉定位准确率达80%以上,为采摘机器人应用于实际生产提供了理论基础。
[Abstract]:It has practical application value to locate the disturbed grape picking point by picking robot in unstructured environment. Firstly, based on the model of "flexible Rod-hinge-rigid Rod-Mass Ball", the shape of grape string is analyzed. The perturbed state is decomposed into two planar pendulum motions of XOY and YOZ. The disturbed grape is photographed by video, and the fruit and fruit stem are obtained by the Otsu threshold segmentation of the multi-frame images of grape in the video. The centroid of grape string in each frame image is calculated, and the curve fitting of the center of grape string in each frame image is carried out to calculate the period and angle of grape single pendulum motion, so as to determine whether the disturbed grape is suitable for visual positioning. For the disturbed grape which can realize visual positioning, the grapevine image corresponding to the centroid point in the middle position of a single pendulum is selected, and the Canny edge detection is carried out on the rectangular area above the grape string. Then the location of the disturbed grape picking point was realized by using Hoff straight line fitting and angle constraint method. The result of visual localization experiment showed that the visual positioning accuracy of the disturbed grape picking point in the natural environment was more than 80% under different illumination. It provides a theoretical basis for the application of picking robot in practical production.
【作者单位】：华南农业大学数学与信息学院;
【基金】：国家自然科学基金项目(31201135) 广东省科技计划项目(2015A020209123) 广州市科技计划项目(201506010081)
【分类号】：TP391.41;TP242

【参考文献】