摘要
随着计算机视觉技术的发展,机器视觉尤其是双目视觉被广泛应用于物体识别、虚拟现实、工业检测、机器人导航和航空航天等领域。图像匹配是机器视觉中的关键技术之一,能否有效地解决该问题严重影响着机器视觉的发展。对现有的匹配算法进行了研究的基础上,针对算法的实时性和精确度提出了改进算法。实验结果表明:误匹配像素百分比与目前常用的基于区域的匹配算法相当,而计算速度却提高了一个数量级,并且边缘特征较好,是一种有效可行的高实时性匹配算法。
With the development of the computer vision, the machine vision especially binocular vision is widely used in many fields such as object recognition, virtual reality, industrial detection, robotic navigation, aviation and spaceflight. Image matching is the one of key techniques in machine vision. The development of machine vision depends on the efficient solution of image matching. Based on study on the existing matching algorithms, an improved algorithm aiming at the real-time and ac- curacy of algorithms is proposed. Experimental results show that the computational speed has been improved for one order of magnitude while the percentage of bad matching Pixels is equivalent to commonly-used area-based matching algorithm. Besides, its edge feature is better, thus it is an effective and feasible stereo matching algorithm with good real-time performance.
出处
《现代电子技术》
2011年第18期46-49,52,共5页
Modern Electronics Technique
关键词
图像匹配
机器视觉
外极线校正
视差图
外极几何
image matching
machine vision
epipolar line rectification
parallax image
epipolar geometry
