摘要
在对比分析了普通环境与阴影环境下图像特点的基础上,提出了一种适用于阴影环境下拖拉机视觉导航的路径识别方法。首先,运用2G-R-B彩色模型分割图像,根据图像的线性灰度分布,采用合理的点运算分析法提高图像对比度,利用迭代阈值分割法和二值图像闭运算提取道路特征;然后,通过扫描道路边缘离散点和最小二乘法拟合出拖拉机的导航路径。实验结果表明,该方法能快速和有效地提高拖拉机视觉导航系统对阴影环境的适应性。
Based on comparing the features of an image in application environment and shadow environment,a new method of road recognition for tractor's vision navigation in shadow environment was proposed in this paper.Firstly,a navigation image was segmented by the 2G-R-B model.Then,by adopting point-operation,iterative-threshold and closing-operation to advance the contrast of the image according to the characters of shadow and pick up the figure of road,and scanning the path points and using least-square,the navigation line of tractor generated.Testing results verified the feasibility and reliability of the method.
出处
《农机化研究》
北大核心
2012年第4期181-184,共5页
Journal of Agricultural Mechanization Research
基金
中国博士后基金项目(20070420190)
江苏省高校自然科学研究计划项目(07KJD460035)
关键词
视觉导航
拖拉机
阴影环境
路径识别
tractor
vision navigation
shadow environment
road recognition