Research on UAV Image Registration Based on Sift Algorithm Acceleration

The paper compares the image registration algorithms: the classical normalized correlation (as a representative of intensity-based algorithms) and the SIFT-based algorithm (feature-based registration). A gradient subpixel correction algorithm was also used for normalized correlation. We compared the effectiveness of their work on real images (including a terrain map) when modeling artificial distortions. The accuracy of determining the position (shift) of one image relative to another in the presence of rotation and scale changes was studied. The experiment was carried out using a simulation model created in the Python programming language using the OpenCV computer vision library. The results of the experiments show that in the absence of rotation and scale changes between the registered images the normalized correlation provides a slightly smaller root-mean-square error. At the same time, if there are even small such distortions, for example, a rotation of more than 2 degrees and a scale change of more than 2 percent, the probability of correct registration for the normalized correlation drops sharply. It was also noted that the advantages of normalized correlation are almost 5 times higher speed and the possibility of using it for small fragments (50x50 or less), where it is problematic for the SIFT algorithm to allocate a sufficient number of keypoints. It was also shown that the use of a two-stage algorithm (SIFT-based registration at the first stage, and optimization with normalized correlation as a criterion at the second) allows you to get both high accuracy and stability to rotation and scale change, but this will be accompanied by high computational costs.

Download Full-text

Fast image registration method based on Harris and SIFT algorithm

Chinese Optics ◽

10.3788/co.20150804.0574 ◽

2015 ◽

Vol 8 (4) ◽

pp. 574-581 ◽

Cited By ~ 3

Author(s):

许佳佳 XU Jia-jia

Keyword(s):

Image Registration ◽

Registration Method ◽

Sift Algorithm

Download Full-text

An Image Registration Algorithm Based on Improved SIFT Feature

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.347-350.3411 ◽

2013 ◽

Vol 347-350 ◽

pp. 3411-3415 ◽

Cited By ~ 1

Author(s):

Yin Wen Dong ◽

Luan Wan ◽

Zhao Ming Shi ◽

Ming Lei Zhu

Keyword(s):

Image Registration ◽

Nearest Neighbor ◽

Interpolation Method ◽

Space Structure ◽

Cruise Missile ◽

Feature Descriptor ◽

Sift Algorithm ◽

Sift Feature ◽

Registration Algorithm ◽

Image Registration Algorithm

Aiming at there are long matching time and many wrong matching in the traditional SIFT algorithm, An image registration algorithm based on improved SIFT feature is put forward. First of all, through setting the number of extreme points in the feature point detection, feature points is found according to the DOG space structure from coarse to fine, and the improved SIFT feature descriptor generation algorithm is used. The preliminary matched point pairs are obtained by the nearest neighbor matching criterion, and the bilateral matching method is used for screening the preliminary matched point. Then, the second matching will be done by the similar measurement method based on mahalanobis distance, and RANSAC algorithm is used to calculate the affine transform model. Finally, the transformed image is resampled and interpolated through the bilinear interpolation method. Experimental results show that the algorithm can realize image registration effectively. Image registration technique is an important research content in computer vision and image processing in the, which are widely used in vehicle matching navigation and positioning, cruise missile terminal guidance, target tracking and recognition, image mosaic[1-6]. SIFT algorithm[3-5]can achieve image registration when there are translation, rotation, affine transformation between two images, even for images took by arbitrary angles. And SIFT feature is the milestone of local feature study. But there are long matching time and many wrong matching in the traditional SIFT algorithm, it is difficult to meet the requirement of fast image registration. This paper puts forward an image registration algorithm based on improved SIFT feature, which is robust for image rotation, affine and scale change, and is better than traditional SIFT algorithm.

Download Full-text