节点文献
基于鱼眼图像的全景漫游模型研究
Researches on Environments of Full View Navigation Based on Fish Eye Image
【作者】 唐俊;
【导师】 吴福朝;
【作者基本信息】 安徽大学 , 计算机应用技术, 2003, 硕士
【摘要】 本文主要讨论了如何利用鱼眼图像建立全景漫游模型的问题。主要工作如下: (1)鱼眼图像的成像原理以及根据其成像原理从鱼眼图像恢复出实际场景图像的算法。 (2)鱼眼图像的轮廓提取的算法。我们给出了一种针对鱼眼图像的特点使用面积统计的方法来计算鱼眼图像的圆形有效区域的圆心和半径的算法,实验证明了我们所给出的方法的正确性和可行性。 (3)对两幅沿相反方向拍摄的鱼眼图像的相似区域进行查找的算法。对图像相似区域的查找一般是用图像匹配的方法,但该方法计算量过大。我们给出了一种通过查找差值图像的极值点来搜索相似区域的方法。该方法计算量较小,而且精度也能达到应用的要求。 (4)鱼眼图像的几何畸变校正。由于鱼眼镜头制造过程中产生的误差等原因,所拍摄的图像几何畸变还是比较严重的,这就产生了对鱼眼图像进行几何校正的问题。我们提出了一种利用球面经纬线模板对图像进行几何校正的方法。 (5)交互式全景漫游的具体实现方法。
【Abstract】 In this paper, we mainly discuss how to use fish eye image to create environments of full view navigation. Our main work is as following:(l)The theory of generating fish eye image and the algorithm for recovering original scene images from fish eye image according this theory.(2)The algorithm for contour retrieval of fish eye image .According the peculiarity of fish eye image, we introduce an algorithm for calculating the center and radius of the circular region in the fish eye image, which use the method of calculating the area of the circular region.(3)The algorithm for searching the similar region of two fish eye image taken along inverse direction. The existing algorithm for searching the similar region is generally image matching, but the computation amount is too much. We design an algorithm for searching the similar region by scanning the extreme value of minus image, which cost less in computation amount and can satisfy the requirement of application in precision.(4) The algorithm for correcting the geometrical distortion of fish eye image. Because of the aberration of fish eye lens resulting from fabrication procedure, there exists serious geometrical distortion, which arises the problem of correcting the fish eye image. We give the algorithm which use spherical template with altitudes and longitudes to correct the image.(5)The algorithm for implementing interactive full view navigation.
【Key words】 Fish eye image; Full view navigation; Contour retrieval; Image matching; Minus image; Geometrical distortion;
- 【网络出版投稿人】 安徽大学 【网络出版年期】2004年 01期
- 【分类号】TP391.41
- 【被引频次】14
- 【下载频次】636