Panoramic image stitching using SIFT
This report describes how SIFT keypoints are obtained by using Difference-of-Gaussian for detecting local maxima and minima and methods to filter detected local features in order to obtain stable keypoints. The magnitudes of gradient and orientations are added to each keypoint before creating each k...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
2009
|
| 主題: | |
| 在線閱讀: | http://hdl.handle.net/10356/17822 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 總結: | This report describes how SIFT keypoints are obtained by using Difference-of-Gaussian for detecting local maxima and minima and methods to filter detected local features in order to obtain stable keypoints. The magnitudes of gradient and orientations are added to each keypoint before creating each keypoint descriptor. These highly distinctive SIFT features are matched against each other to find k nearest-neighbors for each feature. These correspondences are then used to find m candidate matching images for each image. Based on the extracted local features, image transformation matrices are built, images are to be processed based on the transformation matrices. This report also discussed methods for blending images to create seamless panorama. |
|---|