中文摘要：

在月面环境下车轮沉陷严重地影响月球车的移动性能和定位精度,针对现有基于视觉的车轮沉陷检测方法存在系统误差等缺陷,提出了一种基于单应的月球车车轮沉陷视觉测量方法。首先通过分析指出现有的基于视觉的车轮沉陷检测方法存在系统误差。然后引入三维空间中车轮侧面所在平面与摄像机图像平面之间的单应,把由图像处理技术得到的车轮沉陷信息从图像平面变换到车轮侧面所在的平面,在三维空间中计算车轮沉陷,从而消除了现有方法中的系统误差。通过限定感兴趣区域以及引入一维空间滤波器,在图像中检测车轮沉陷信息更加快速、鲁棒。仿真数据实验以及月球车原理样机在模拟月面地形下的真实图像数据实验都表明新方法是有效可行的。

英文摘要：

The wheel sinkage greatly affects the mobility and localization accuracy of a lunar rover in deformable lunar terrains.To overcome the deficiency of existing methods,a novel vision-based method is presented to measure the sinkage.Firstly it is proved that there are system errors in the existing methods.Then the homography between the wheel plane in 3D space and the image plane of a mounted camera is introduced,and the wheel sinkage information detected by image processing techniques are transformed from the image plane into the wheel plane.Finally the wheel sinkage is computed in the wheel plane.Contact points can be rapidly and robustly detected in the image by narrowing the interesting area and introducing a one-dimensional space filter.The proposed method can eliminate system errors introduced by existing methods.Results of experiments show its validity and feasibility.