To alleviate the effect of matching of corresponding feature points and extraction error of single feature point, a method by using a single camera as monocular measurement is presented based on image processing. Firstly, this paper sets up the mapping relationship between image point and target point, and establishes pinhole imaging model. Secondly, it describes the mapping relationship between object area and image area of object by using image analysis, and establishes the model of distance measurement in optical direction. Then, the principle of distance measurement between optical center and feature point is proposed after image processing and feature point extracting are carried out. At last, the paper starts out verification experiments and analyzes the cause that error increases with further distance. After analyzing the data of experiment, the conclusion that error is related with the optical axis deviation is made. Sequentially, with the maximum relative error 1.68％ of revised data, there is a remarkable improvement which proves the feasibility and the effectiveness of the proposed principle.%为了克服对应点匹配和单个特征点提取误差对测量结果的影响,本文在基于图像处理的基础上,提出了一种基于特征点的单目视觉测距方法.首先利用小孔成像原理,得出成像点与目标点的映射关系,建立小孔成像模型.然后通过对目标图像的分析,得出目标物与目标图像的面积映射关系,建立视觉测量的直线测距模型;通过图像处理,提取目标图像的特征点,将光心与目标物的距离关系转化为光心与特征点的距离关系,提出了基于特征点的单目视觉测距原理.最后对测距原理进行实验验证并对得出的数据进行分析,针对测量误差随距离增加而变大的现象,通过对数据的分析,得出是由于光轴偏心而引起的误差;将数据修正后,测量结果的最大相对误差为1.68%,精度提高明显,验证了原理的可行性和算法的有效性.
展开▼
