Improved SIFT Algorithm Research Articles

Improved SIFT image registration algorithm on characteristic statistical distributions and consistency constraint

In the process of SIFT (Scale Invariant Feature Transform) image registration algorithm; the principal orientation is affected by the dispersion of histogram. Besides; the feature descriptor section of conventional sift does not make full use of local feature information. As to these problems; an improved sift algorithm on characteristic statistical Distributions and consistency constraint will be presented in this paper. Firstly; DoG scale space feature point detection method is adopted to extract key points. Then; in the process of principal orientation generation; our method selects line with maximum dispersion. Furthermore; this method generates feature descriptor based on characteristic statistical Distributions in polar coordinate. Finally; a new matching method based on consistency constraint will be introduced. In experiments; we test the performances of our propose method. The experimental results demonstrate the feasibility and effectiveness of our approach.

Image Mosaic Method Based on Gaussian Second-order Difference Feature Operator

<p><em>To compose the wide visual angle and high resolution image from the sequence of images which have overlapping region in the same scene quickly and correctly, an improved SIFT algorithm which is based on D2oG interest point detector was proposed. It extracted the image feature points and generated corresponding feature descriptors by improved SIFT algorithm. Then, using the random consistency (RANSAC) algorithm purified feature point matching pairs and calculating the transformation matrix H. Last, complete the seamless mosaic of images by using the image fusion algorithm of slipping into and out. It respectively process the images which had the four typical transformations with the traditional SIFT and the proposed method. The result indicated that the number of feature pairs is fewer than SIFT algorithm and the mosaic time is shorter, and then the matching efficiency is higher than the later. This proposed method reduces the complexity of operation and improves real-time of image mosaic simultaneously.</em></p>

Image Mosaic Method Based on Gaussian Second-order Difference Feature Operator

To compose the wide visual angle and high resolution image from the sequence of images which have overlapping region in the same scene quickly and correctly, an improved SIFT algorithm which is based on D2oG interest point detector was proposed. It extracted the image feature points and generated corresponding feature descriptors by improved SIFT algorithm. Then, using the random consistency (RANSAC) algorithm purified feature point matching pairs and calculating the transformation matrix H. Last, complete the seamless mosaic of images by using the image fusion algorithm of slipping into and out. It respectively process the images which had the four typical transformations with the traditional SIFT and the proposed method. The result indicated that the number of feature pairs is fewer than SIFT algorithm and the mosaic time is shorter, and then the matching efficiency is higher than the later. This proposed method reduces the complexity of operation and improves real-time of image mosaic simultaneously.

An automatic optical and SAR image registration method with iterative level set segmentation and SIFT

Although optical image registration methods have been successfully developed over the past decades, the registration of optical and synthetic aperture radar (SAR) images is still a challenging problem in remote sensing. Feature-based methods are considered to be more effective for multi-source image registration. However, almost all of these methods rely on the feature extraction algorithms. In this article, a simultaneous segmentation and feature-based registration method based on an iterative level set and scale-invariant feature transform (ILS-SIFT) is proposed. The core idea consists of three aspects: (1) an iterative procedure that combines image segmentation and matching is proposed to avoid registration failure caused by poor feature extraction; (2) a uniform level set segmentation model for optical and SAR images is presented to segment conjugate features; and (3) an improved SIFT algorithm is employed to determine whether the registration was successful. Experimental results have shown the effectiveness and universality of the proposed method.

Research on Disparity Map Generation Method of Underwater Target Based on the Improved SIFT Algorithm

In order to obtain the depth information of the underwater target, it’s necessary to generate the disparity map based on binocular vision stereo matching. In the circulation water channel, the stereo matching experiments with underwater target were carried out by using the BM algorithm, SGBM algorithms and SIFT algorithm respectively. Then the characteristics of the disparity maps were analyzed for the three kinds of stereo matching algorithms. Compared with the BM algorithm and SGBM algorithms, the SIFT algorithm has been proved to be more suitable for underwater stereo matching. In order to obtain more feature points of underwater image, it is necessary to improved SIFT algorithm parameter. Underwater image matching experiments were made to determine the principal curvature coefficientγ. The results illustrated that the improvedγis better than the original value for underwater disparity map generation.

Improved SIFT Algorithm Based on Mixed Programming

In order to improve the generality and real-time of image matching procedure, Visual Studio 2010 and MATLAB R2009a have been used as the platform to research mixed programming and improved SIFT algorithm. In this method, the advantages of C # and Matlab have been combined to reduce the difficulty of programming and to improve programming efficiency. The results show that, improved SIFT algorithm can greatly improve real-time of matching program while guaranteeing good matching rate, its suitable in real-time applications.

Parallel Research and Implementation of SAR Image Registration Based on Optimized SIFT

A new SAR image registration method was Proposed based on improved SIFT algorithm. Which adopted multi-core system platform was used to overcoming the problem of high complexity algorithm of SIFT algorithm; According to the characteristics of SAR image, first of all, the source SAR image was enhanced in airspace, and finish the parallel extraction of feature points with the improved SIFT algorithm, then used Euclidean distance and the RANSAC algorithm to complete the matching of feature points and eliminate unmatching, finally realizes the SAR image registration. The experimental results show that the method can guarantee in the registration precision and reduce the complexity of the registration. DOI : http://dx.doi.org/10.11591/telkomnika.v12i2.4414

A Novel Image Registration Approach With SIFT Algorithm and Tangent-Cross-Point Feature

By the aid of Tangent-Cross-Point feature, a novel improved SIFT algorithm is proposed in this paper. This method is more efficient and reliable in some cases, since it adopts Tangent-Cross-Point feature to improve the efficiency and robustness of SIFT algorithm. With a pair of matching feature points and a pair of curves nearby, it’s very easy to estimate the values of local transformation parameters and validate the correctness of this matching by the aid of Tangent-Cross-Point feature.

A New Algorithm of Fast-Generated Panoramic Images

The new algorithm of fast-generated panoramic images this paper puts forward is to extract the feature points of images by the improved SIFT algorithm, and use Euclidean distance combining the K-D tree structure to realize the rapid initial feature matching. Then, based on these initial matching points and the theory of random sampling consistent algorithm, the purification of feature points is realized. At last, the introduction of correction coefficient makes it possible to eliminate fusion ghosts, and HIS space image fusion is applied in order to eliminate the brightness differences. It is verified by the experiments that on the premise of generation of quality guarantee, the new algorithm greatly improves the generation efficiency of panorama images.

Adaptive SIFT Matching Relay Tracking of Variable Scene based on Intelligent Agent

The goal of continuous tracking between different video problem, a change of scene with Intelligent Agent between the goal of achieving the automatic extraction and identification; improved SIFT algorithm, to factor in the selection of scale space with the goal of scaling the size of adaptive change, overcome image rotation, scale Scaling, brightness change, disturbance, maintaining a change of perspective, affine transformation, noise stability, from different imaging conditions and complex scenes of dynamic video images to extract key points and matched to achieve different video continuous track between the target image. experiments show that the algorithm real time, robustness, matching accuracy.

Image Target Identification of UAV Based on SIFT

The paper finds a method to meet the large amounts of information and real-time processing of the UAV(Unmanned Aerial Vehicle) image, it could increase efficiency of target identification. The SIFT algorithm has good robustness, it could overcome some affect of image deformation and block, but it could not meet the real-time processing of UAV image. The paper use simplified Forstner operator to improve SIFT algorithm, reduce the computation of feature point recognition. Through the simulation, we prove that the improved SIFT algorithm could meet the accuracy and speed requirements of UAV image in the complex background.