节点文献
基于形状模板匹配的图像拼接算法
An Image Mosaic Algorithm Based on Matching Shape Template Metric
【摘要】 文章提出了一种基于形状模板匹配的图像自动拼接方法。提取图像的角点作为特征点,利用归一化梯度模板对其进行预匹配,然后利用形状模板在四个方向对模板内图像的边缘点与模板边界的最短距离进行统计,获取模板图像的结构特征向量以实现对特征点的精确匹配。实验结果表明该算法具有较好的实用价值。
【Abstract】 In this paper,a novel image mosaic algorithm is proposed.Firstly,extract the corner points of the image as features and use a gradient template to find the matching points roughly.Then use a shape template in four directions,to measure the distance from each window boundary point to the nearest point of the edges in the template.By this scheme,we can get the shape feature vector of the template and use them to match the feature points accurately.The experiments show this algorithm performs well in the application of image mosaic.
【关键词】 梯度模板;
形状模板;
图像拼接;
全景图;
【Key words】 gradient template metric; shape template metric; image mosaic; panorama;
【Key words】 gradient template metric; shape template metric; image mosaic; panorama;
【基金】 山东省自然科学基金重点资助项目(编号:C2005G02);山东省优秀中青年奖励基金资助项目(编号:03BS001)
- 【文献出处】 计算机工程与应用 ,Computer Engineering and Applications , 编辑部邮箱 ,2006年28期
- 【分类号】TP391.41
- 【被引频次】15
- 【下载频次】529