基于双目立体视觉的规则物体体积测量方案  

Volume Measurement Scheme of Regular Object Based on Binocular Stereo Vision

作　　者：吴琪波 WU Qibo(School of Aerospace Engineering,Xiamen University,Xiamen 361102,China)

机构地区：[1]厦门大学航空航天学院,福建厦门361102

出　　处：《现代信息科技》2020年第18期112-114,118,共4页Modern Information Technology

摘　　要：基于双目立体视觉与图像处理,文章提出一种对于长方体体积测量的方案。首先对双目相机进行标定,拍摄目标物体得到左右两张图像,通过半全局匹配算法进行图像匹配,从而得到视差以及深度信息。将图片的深度信息转化为灰度信息并对灰度图进行处理,提取出目标的上表面外接矩形面积以及相对于地面的高度信息,从而得到目标物体积。实验结果表明,提出的体积测量方法在一定范围内达到了精度标准。This paper presents a scheme for cuboid volume measurement based on binocular stereo vision and image processing.First,the binocular camera was calibrated,and the target was photographed to obtain the left image and right image.Semi-global matching algorithm was used for image matching to obtain the disparity and depth information.The depth information of the image is converted into gray information.The gray image is processed to extract the upper surface area of the target and the height information relative to the ground,and the volume of the target can be get.The experiments show that this volume measurement method can be effect within a certain range.

关 键 词：双目立体视觉 立体匹配 图像处理 体积测量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]