基于边缘特征点的全景图像拼接算法

Panoramic Image Mosaic Algorithm Based on Edge Feature Points

全景图像拼接技术已经成为图像处理领域中的一个研究热点.传统的基于特征的全景图像拼接方法存在旋转、扭曲对匹配结果影响较大的缺点,拼接质量不高.为改善和提高拼接效果和扩展度,首先对图像边缘特征点的配准、几何坐标优化、空间坐标变换、最优投影变换参数的求解、融合等方面展开研究,然后基于此改进相应算法,提出一种基于边缘特征点的全景图像拼接算法,并进行了应用实验.实验结果表明,所给算法比传统算法具有更良好的拼接效果和实用价值.

It is known to all that the panorama mosaic technology has become a research hotspot in the field of the image processing. The traditional panorama mosaic methods based on features have disadvantages for twist and rotation to big influence on the matching result, so that the mosaic quality is not high apparently. To improve and enhance the patchwork effect and divergence, firstly, the research is done on the solution and fusion of the image edge feature point registration, geometric coordinate optimization, space coordinate transformation, optimal projection transformation. Then, a type of panoramic image mosaic based on its improved related algorithm is put forward and put into the application experiment. It turns out that the proposed algorithm is much better in patchwork effect and practical value than the traditional algorithm.