基于图像的大型物体尺寸的三维测量方法研究  

Research of 3D Large Object Size Based on Measurement Method of Image

作　　者：师蒙蒙[1] 韩焱[1] 王鉴[1] 刘鹏[1] 杨海[1] 

机构地区：[1]中北大学电子测试技术国家重点实验室,山西太原030051

出　　处：《计算机测量与控制》2011年第10期2351-2353,共3页Computer Measurement &Control

基　　金：山西省自然基金项目(2007012003)

摘　　要：针对大型物体尺寸不易测量的问题,研究简单快速的测量方法;提出了一种利用数码相机完成物体尺寸的三维测量的方法;利用数码相机从多角度获取图像,采用稳定性强的Harris角点检测算法对图像进行特征点检测;在相机标定过程中提出了一种利用图像中被测物体自适应寻找标定模板的方法,扩大了标定模板的使用范围;最后利用投影矩阵理论计算特征点的世界坐标进行测量;实验证明测量结果误差在2.0%以内,说明该方法能够克服传统测量方法的缺点,满足一般工业的测量要求并且适用范围广。In allusion to the problem that the large object size is difficult to measure, this paper researches simple and quick measuring methods and presents a method of 3D measurement of object＇ s size based on image using digital camera. Firstly, we obtain images from multi angles using a digital camera and detect characteristic points using Harris corner detection algorithm which is strong stability. Secondly, a new method of adaptive looking for calibration template based on the object to be tested in images is proposed in the processing of the calibration, and expands the range of application of the calibration target. Finally, we compute the world coordinates of the characteristic points to measure using projection matrix theory. Experiments prove that measuring results is within two percent, which show that this method can overcome the disadvantages of the traditional measurement and meet requirements of genera/industrial measurement and application scope is wide.

关 键 词：三维测量 角点检测 标定模板 投影矩阵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]