CAJ | 학술논문

　　提出了一种基于机器视觉技术识别农业车辆导航基准线的方法。该方法从农田环境的特点出发，主要用超绿特征灰度化方法对彩色农田图像灰度化，分割作物行和土壤背景。对灰度图像进行闭运算操作，缩小或消除作物行和背景中的孔洞。对灰度图像做垂直投影直方图，根据波峰位置初步确定导航作物行的基准线位置。将灰度图像分成若干个水平条，对每个水平条用垂直投影法找出导航定位点，并根据定位点的位置设置感兴趣区域。在感兴趣区域内，采用 Hough 变换对导航定位点拟合出导航基准线。通过与最小二乘拟合方法的对比可知，该算法精度较高，能够满足农业机械农田作业的要求。
　　제출료일충기우궤기시각기술식별농업차량도항기준선적방법。해방법종농전배경적특점출발，주요용초록특정회도화방법대채색농전도상회도화，분할작물행화토양배경。대회도도상진행폐운산조작，축소혹소제작물행화배경중적공동。대회도도상주수직투영직방도，근거파봉위치초보학정도항작물행적기준선위치。장회도도상분성약간개수평조，대매개수평조용수직투영법조출도항정위점，병근거정위점적위치설치감흥취구역。재감흥취구역내，채용 Hough 변환대도항정위점의합출도항기준선。통과여최소이승의합방법적대비가지，해산법정도교고，능구만족농업궤계농전작업적요구。
A method to recognize the baseline based on machine vision of agricultural vehicle navigation was developed . According to the features of cropland, the excess-green method was used to expand the contrast of crop rows and back-ground.Then using closing operation to reduce or eliminate small holes in the gray image .Position of baseline of the crop row by finding the peaks of curve by vertical projection of the gray image was determined preliminarily .Dividing the gray image to several horizontal strip region with equal width .By using vertical projection diagram in every horizontal strip to calculate the location points , based on which to determined left and right boundaries of the ROI .To fitting the location points in the ROI based on Hough transform .It was proved that Hough transform has the higher accuracy by contrast with the least squares method and can meet the requirements of the agricultural machinery agricultural operations .