﻿ 基于对SIFT算法优化的图像拼接技术 Image Mosaic Technology Based on an Optimized Method of SIFT Algorithm

Journal of Image and Signal Processing
Vol.04 No.04(2015), Article ID:16232,7 pages
10.12677/JISP.2015.44017

Image Mosaic Technology Based on an Optimized Method of SIFT Algorithm

Cheng Yang, Xiaogang Xu, Guanlei Xu

Email: 245591167@qq.com

Received: Oct. 6th, 2015; accepted: Oct. 21st, 2015; published: Oct. 27th, 2015

ABSTRACT

In order to solve the problems of wrong feature points in the process of image mosaic algorithm based on SIFT, an optimized method is put forward. Using a new algorithm to eliminate the wrong feature points, and using the least square method to fit the new picture not only improve the matching accuracy, but also improve the image mosaic result and reduce the time of the image mosaic. Experimental results demonstrate that the method can produce high quality image mosaic.

Keywords:SIFT, Image Mosaic, The Least Square Method

Email: 245591167@qq.com

1. 引言

2. SIFT算法的特征匹配

2.1. 构建尺度空间

2.2. 点的搜索与定位

Figure 1. DOG scale-space

Figure 2. Detects key points in DOG scale-space

(1)

(2)

2.3. 赋予特征点128维参数方向

2.4. 特征描述

3. 利用Ransac算法进行特征点筛选

3.1. Ransac算法的描述

1) 内外点的距离阈值。这个阈值用来判定数据点是内点或者外点。

2) 估计次数N。令w为数据是真是数学模型内点的概率，n为确定模型参数的最少点数，如果要确定变换关系，则需要4个匹配对，即n = 4。这样，依次估计中使用的所有n个点都为内点的概率就是。如果要保证经过N次至少有一次估计中的所有数据点都是内点的概率是p，那么N需要满足：

3) 一致性集合大小阈值。估计得到模型的参数后，还要统计整个数据集中符合该模型的内点数目t。令数据点是错误模型内点的概率为y，我们希望错误模型的内点数目t越小越好，即

3.2. 本文的方法

Figure 3. The gradients information of key points generate eigenvector

Figure 4. Schematic diagram of the least square method

4. 实验结果

5. 结束语

(a) (b) (c) (d)(e) (f)

Figure 5. (a) Mosaic image 1; (b) Mosaic image 2; (c) SIFT algorithm matching; (d) Optimized matching; (e) Mosaic results; (f) Optimized mosaic results

Table 1. Comparison between SIFT and optimized method

Image Mosaic Technology Based on an Optimized Method of SIFT Algorithm[J]. 图像与信号处理, 2015, 04(04): 139-145. http://dx.doi.org/10.12677/JISP.2015.44017

1. 1. 袁杰 (2013) 基于SIFT的图像配准与拼接技术研究. 南京理工大学, 南京.

2. 2. Moravec, H.P. (1981) Rover visual obstacle avoidance. International Joint Conference on Artificial Intelligence, 3, 117-118.

3. 3. Harris, C.G. and Stephens, M. (1988) A combined corner and edge detector. Proceedings of Fourth Alvey Vision Conference, 147-151.

4. 4. Lowe, D.G. (2004) Distinctive image features from scale-invariant key points. International Journal of Computer Vision, 60, 91-110. http://dx.doi.org/10.1023/B:VISI.0000029664.99615.94

5. 5. Perona, P., Malik, J., Perona, P., et al. (1990) Scale-space filtering and edge detection using anisotropic diffusion. IEEE Transactions on Pattern Analysis & Machine Intelligence, 12, 629-639. http://dx.doi.org/10.1109/34.56205

6. 6. Lindeberg, T. (1994) Scale-space theory: A basic tool for analysing structures at different scales. Journal of Applied Statistics, 21, 224- 270.

7. 7. Liao, K., Liu, G. and Hui, Y. (2013) An improvement to the SIFT descriptor for image representation and matching. Pattern Recognition Letters, 34, 1211-1220. http://dx.doi.org/10.1016/j.patrec.2013.03.021

8. 8. Zhong, S., Wang, J., Yan, L., et al. (2013) A real-time embedded architecture for SIFT. Journal of Systems Architecture the Euromicro Journal, 59, 16-29. http://dx.doi.org/10.1016/j.sysarc.2012.09.002

9. 9. Wu, X., Zhao, Q. and Bu, W. (2014) A SIFT-based contactless palmprint verification approach using iterative RANSAC and local palmprint descriptors. Pattern Recognition, 47, 3314-3326. http://dx.doi.org/10.1016/j.patcog.2014.04.008