计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (22): 191-194.

• 图形、图像、模式识别 • 上一篇    下一篇

一种基于平均垄间距的视觉导航垄线识别算法

张志斌,潘华稳,李 琛,王冰清   

  1. 内蒙古大学 计算机学院,呼和浩特 010021
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-08-01 发布日期:2011-08-01

Crop rows identification based-row interval for field vision guidance system

ZHANG Zhibin,PAN Huawen,LI Chen,WANG Bingqing   

  1. College of Computing,Inner Mongolia University,Hohhot 010021,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-08-01 Published:2011-08-01

摘要: 通过对经过灰度二值化处理后的田间农作物图像进行基于平均垄间距行向连续区域中点提取(行提取)和基于平均垄间距列向最近邻搜索目标点提取(列提取),实现农作物图像垄行结构的识别。行提取为基于垄间距一致性特点,通过连续区域聚类求中点以消除大面积杂草等噪声影响;列提取为基于垄列向的连续性,在行提取的基础上对所得中点在垄间距范围进行最近邻搜索,以减少非垄孤立点干扰因素的影响。通过最小二乘法拟合获得各垄线段,以克服失垄、断垄等田间自然光照条件下对垄线图像识别的影响。实验1采用30幅单垄图像,垄行方向角平均误差为-0.66°;实验2采用20幅多垄图像,垄行方向角平均误差为1.31°,最长时耗约为6 ms,满足田间视觉导航系统实时性、准确性要求。

关键词: 作物垄行, 图像处理, 最邻近搜索, 最小二乘法拟合, 视觉导航

Abstract: A crop rows identification based-row interval is proposed in this paper including three main procedures.The row scanning firstly is implemented to decrease the effect of great area grass.The column scanning is implemented to find out the crop points based on the nearest neighbor search,in which the non-crop isolated points in image are deleted.The least-square method is used to fit the crop lines with high robustness for the losing crop,or part crop problems.Experimental results show that the algorithm proposed in this paper is efficient,with high real-timeliness,the time consumption being at most 6 ms,and crop row angle error is -0.66° for experiment one,1.31° for experiment two.

Key words: crop rows, image processing, nearest neighbor search, least-square fitting, vision navigation