农机化研究2025，Vol.47Issue(9)：22-27,34,7.DOI:10.13427/j.issn.1003-188X.2025.09.004
基于机器视觉的马铃薯收获导航线检测方法
Detection Method of Navigation Line of Potato Harvesting Based on Machine Vision
摘要
Abstract
In order to accurately and quickly extract the dividing line between potato harvest area and non-harvest area,a navigation line detection method based on improved vertical projection method was proposed.The super green feature algo-rithm(ExG)was used to extract the green feature of potato stem and leaf in real time,and the Otsu binarization process was carried out.The image denoising method based on area was used to denoise the image.The image after noise removal was divided horizontally,and the vertical projection curve of pixels was analyzed according to each horizontal bar in turn.The peak point detection algorithm was used to extract the potato dividing line feature points.For the non-first frame im-age,the"reference line"was introduced,and the distance between each feature point and the reference line was com-pared with the distance threshold to eliminate the points that deviate far from the reference line.The least square method was used to fit the remaining feature points to obtain the potato boundary line.The experimental results showed that the al-gorithm had high robustness,the extraction accuracy was more than 92%,and the average time is about 140ms,which can realize the accurate extraction of potato dividing line,and provide reliable navigation information for the automatic navigation of potato harvest.关键词
机器视觉/导航线/水平条划分/垂直投影/最小二乘法/马铃薯Key words
machine vision/navigation line/horizontal bar division/vertical projection/least square method/potato分类
计算机与自动化引用本文复制引用
潘良,曹中华,董继伟,李亚丽,湛小梅,刘汶树,李果..基于机器视觉的马铃薯收获导航线检测方法[J].农机化研究,2025,47(9):22-27,34,7.基金项目
重庆市科研机构绩效激励引导专项(cstc2022jxjl80008) （cstc2022jxjl80008）
重庆市农业科学院绩效激励引导专项(cqaas2021jxj133) （cqaas2021jxj133）
