【作者】 李井辉；

【导师】 唐国维；

【作者基本信息】 大庆石油学院 ， 计算机应用技术， 2004， 硕士

【摘要】 作为一个多学科交叉的领域，计算机视觉的理论研究和实际应用都取得了飞速的发展。尤其是近年来，图像科学的发展和计算机信息处理能力的增强，为计算机视觉的研究和应用提供了良好的条件。因此，对计算机视觉的进一步研究，具有重要的理论意义和实用价值。 计算机立体视觉的主要研究内容是由多幅二维的平面图像恢复出被摄物体的三维坐标，而其中基于两幅图像的双目视觉技术则是一个研究热点。双目立体视觉的基本原理是模仿人眼与人类视觉的立体感知过程，从两个视点观察同一景物，以获取不同视角下的感知图像，通过三角测量原理计算图像像素间的位置偏差，以获取景物的三维信息。 一个完整的双目视觉系统通常可分为图像获取、摄像机标定、特征提取、立体匹配、深度恢复和深度插值等六大部分。本论文主要是图像特征提取、立体匹配这两方面的内容展开研究。首先根据选择的立体视觉成像系统模型阐述了立体视觉测距的基本原理。其次，对图像特征提取等预处理方法进行详尽地分析，通过实例对算法进行实验。同时，对图像匹配问题涉及的内容包括图像匹配方式、质量控制策略等进行了总结；最后，以特征提取与图像匹配研究为基础，针对不同类型的立体图像提出了基于图像边缘特征和图像兴趣点的匹配方法，并通过实例对算法进行验证和分析，取得了较好的匹配效果。更多还原

【Abstract】 As a new field based on various interdisciplinary, the theory research and practical application of the Computer Vision make great progress. Especially in recent years, the development of image processing technology and high performance computer coming into being provide favorable conditions for computer vision research and application. So. the further research to computer vision have much theory meaning and practicality value.Stereo vision is to reconstruct the three-dimensional coordinate of an object from some of its two-dimensional images. Generally, one would like to reconstruct the object in three-dimensional space with only two different view points. The fundamental principle of computer vision with dual-cameras have a strong resemblance to the visual perception procedure of mankind. We can get different images with two different view points and then calculate the position deviation of images pixels based on triangular measure. Finally we can reconstruct the three-dimensional coordinate of the object.The integrated binocular stereo vision system generally includes image acquisition, camera calibration, features extraction stereo matching, depth recovery and depth interpolation, etc. The features extraction and the stereo matching is mainly discussed in this paper. First, we elaborate the fundamental principle of measure distance in stereo vision based on camera modeling. Second, the paper gives the elaborate analysis on preprocessing methods such as features extraction, and vrifies the algorithms by example. Furthermore, it summarizes the content of stereo matching such as match method, the strategy of match quality control. Finally, we submit the match algorithm based on image edge features and match algorithm based on the image interesting dot by features extraction and image matching. Two kinds of algorithm is in accordance with each images, but we get better match results by experiments.更多还原