图像处理在猕猴桃采摘机器人路径规划中的应用

为了实现猕猴桃采摘自动化,设计了采摘机器人导航行走系统。该系统基于图像处理,以种植垄和杂草空地分界线作为导航轨迹,主要包括以下内容:①图像处理,采用HSV系统中色调H分量,实现图像灰度处理,采用同态滤波和阈值分割,实现图像二值化,最终找到导航轨迹离散点;②对离散导航像素进行拟合,并将图像中导航方程转化到实际空间坐标系中,得到实际导航方程;③对系统导航精度和抗光照干扰性能测试,表明系统能有效抑制光强变化带来的影响。系统结构简单,性能可靠,适宜大范围推广。

The Appliation of Image Processing in Path Planning for Kiwifruit Picking Robot

Inorder to achieving automation picking for kiwifruit,navigation running system for picking robot was designed.This system based on image processing.Dividing line between column of kiwifruit and open-field weeds was taken as navigation trajectory.This system composed of following parts:1)Image processing,The hue H in HSVsystem was used to process gray image.Binary image was achieved by homomorphic filtering and threshold segmentation.The discrete points of navigation trajectory was found.2)Linear fitting for discrete points of navigation trajectory,coordinate transformation was finished,and navigation equation in space coordinate system was achieved.3)Test for navigation accuracy and light interference were taken,the result showen that,this system had high precision,and inhibitted light interference effect.This system with simple structure and reliable performance,can be popularized.