论文部分内容阅读
讨论了用于地面自主车(ALV)视觉导航的红外道路图像的特点,论述了红外道路图像的预处理策略和方法,分析了一些常用的图像处理方法在处理红外道路图像时的缺陷,指出由于温差的渐变性,红外道路图像的路边检测特别适合使用阈值穷举法,并给出了在温度渐变的路边上边缘强度的求取方法和使用二阶微分算子得到边界线段的精确位置的方法。同时使用了边缘信息和灰度信息来提取道路边界信息,阈值穷举法和二阶微分算子得到的道路边界线段相互融合,得到候选的道路边界线段,在ALV的实际应用中取得了很好的效果。
This paper discusses the characteristics of infrared road image used in ALV visual navigation, discusses the preprocessing strategy and method of infrared road image, analyzes some defects of common image processing methods when dealing with infrared road image, The gradients of temperature difference and roadside detection of infrared road images are especially suitable for using the threshold value exhaustive method, and the method of obtaining the edge strength on the roadside with temperature gradient and the exact position of the boundary line segment obtained by using the second derivative operator Methods. At the same time, edge information and grayscale information are used to extract the information of road boundary, and the threshold value exhaustive method and the second-order differential operator are used to fuse the road boundary line segments to get the candidate road boundary line segment, which is very good in the practical application of ALV Effect.