摘要
从实际应用角度,对大视差稠密视差图误匹配滤波方法进行了研究,给出了一种针对前景区域进行误匹配滤波的邻域视差相关性三维空间滤波算法。实验结果表明,该算法能够明显地去除并修正视差图的误匹配点,提高视差图质量。
Proceeding from practical application, the false match filering method of large-parallax dense image is researched. A three-dimensional space filtering algorithm which aims at foreground areas to filtrate false matches based on neighborhood parallax relativity is designed and realized. Experiments show that the method could eliminate and correct false match points of parallax images and improve the quality of parallax image.
出处
《物联网技术》
2011年第2期63-65,共3页
Internet of things technologies
关键词
立体视觉
稠密视差图
误匹配滤波
三维空间滤波法
Three-dimensional perspective
dense parallax
false match filtering
three-dimensional space filtering method
