基于视觉的茶作物行间行走路径规划研究

Research of Tea Garden Path Planning Based on Machine Vision

以茶作物为对象,提出了一种基于视觉的作物行间行走路径规划方法。对实时彩色作物图像用2G-RB算法灰度化,分离作物和背景,通过固定阈值二值化图像,采用形态运算去除干扰。以区域内作物的比例变化为判别垄间垄头依据。对于垄间,利用漫水填充算法分离出垄沟,通过最小二乘法拟合导航线,得到垄间行走时候的导航控制参数;对于垄头,先分离出转弯一侧的作物,采用圆形模型拟合垄头处轮廓,通过拟合的圆来规划出转弯时候路径。试验结果表明:该算法能准确提取导航线,处理一帧图像约为150ms左右,可满足导航的要求。

A method of path-planning based on machine vision for tea crop walking is proposed in this paper. After griz- zled with the 2G - R - B processing, the tea crop in the color images is distinguished from background. The binary im- age with expectations is obtained by the fixed threshold and morphological operations. By calculating the proportion change of tea plants in the fixed areas, the location of agricultural machine can be determined. If it is in the middle of crop line, the navigation area can be segmented out with flood water filling algorithm to get navigation discrete points, with which the navigation line is fitted through the method of least squares. If it is at the end of crop line, the edge of crop segmented out by flood water filling algorithm on the turning side is extracted, the route is planned through the circle fitted on the basis of the edge. Experimental results show that the proposed method takes about 150ms to obtain the navi- gation line. It can satisfy the requirements of real-time and accuracy of navigation.