Sub-pixel Matching Research Articles

QUANTITATIVE ASSESSMENT OF A NOVEL SUPER-RESOLUTION RESTORATION TECHNIQUE USING HiRISE WITH NAVCAM IMAGES: HOW MUCH RESOLUTION ENHANCEMENT IS POSSIBLE FROM REPEAT-PASS OBSERVATIONS

Higher spatial resolution imaging data is always desirable to the international community of planetary scientists interested in improving understanding of surface formation processes. We have previously developed a novel Super-resolution restoration (SRR) technique (Tao & Muller, 2016) using Gotcha sub-pixel matching, orthorectification, and segmented 4th order PDE-TV, called GPT SRR, which is able to restore 5 cm-12.5 cm near rover scale images (equivalent to Navcam projected FoV at a range of ≥ 5 m) from multiple 25 cm resolution NASA MRO HiRISE images. The SRR technique has been successfully applied to the rover traverses for the MER and MSL missions within the EU FP-7 PRoViDE project. These SRR results have revealed new surface information including the imaging of individual rocks (diameter ≥ 25 cm) by comparison of the original HiRISE image and rover Navcam orthorectified image mosaics. In this work, we seek evidence from processing a very large number of stereo reconstruction results from all Navcam stereo images within PRoViDE, registration and comparison with the corresponding SRR image, in order to derive a quantitative assessment on key features including rocks (diameter < 150 cm) and rover track wheel spacing. We summarise statistics from SRR-Navcam measurements and demonstrate that our unique SRR datasets will greatly support the geological and morphological analysis and monitoring of Martian surface and can also be applied to landing site selection, in order to avoid unsuitable terrain, for any future lander/rover as well as help to define future rover paths.

A NOVEL SUB-PIXEL MATCHING ALGORITHM BASED ON PHASE CORRELATION USING PEAK CALCULATION

The matching accuracy of homonymy points of stereo images is a key point in the development of photogrammetry, which influences the geometrical accuracy of the image products. This paper presents a novel sub-pixel matching method phase correlation using peak calculation to improve the matching accuracy. The peak theoretic centre that means to sub-pixel deviation can be acquired by Peak Calculation (PC) according to inherent geometrical relationship, which is generated by inverse normalized cross-power spectrum, and the mismatching points are rejected by two strategies: window constraint, which is designed by matching window and geometric constraint, and correlation coefficient, which is effective for satellite images used for mismatching points removing. After above, a lot of high-precise homonymy points can be left. Lastly, three experiments are taken to verify the accuracy and efficiency of the presented method. Excellent results show that the presented method is better than traditional phase correlation matching methods based on surface fitting in these aspects of accuracy and efficiency, and the accuracy of the proposed phase correlation matching algorithm can reach 0.1 pixel with a higher calculation efficiency.

A NOVEL SUB-PIXEL MATCHING ALGORITHM BASED ON PHASE CORRELATION USING PEAK CALCULATION

Abstract. The matching accuracy of homonymy points of stereo images is a key point in the development of photogrammetry, which influences the geometrical accuracy of the image products. This paper presents a novel sub-pixel matching method phase correlation using peak calculation to improve the matching accuracy. The peak theoretic centre that means to sub-pixel deviation can be acquired by Peak Calculation (PC) according to inherent geometrical relationship, which is generated by inverse normalized cross-power spectrum, and the mismatching points are rejected by two strategies: window constraint, which is designed by matching window and geometric constraint, and correlation coefficient, which is effective for satellite images used for mismatching points removing. After above, a lot of high-precise homonymy points can be left. Lastly, three experiments are taken to verify the accuracy and efficiency of the presented method. Excellent results show that the presented method is better than traditional phase correlation matching methods based on surface fitting in these aspects of accuracy and efficiency, and the accuracy of the proposed phase correlation matching algorithm can reach 0.1 pixel with a higher calculation efficiency.

A Novel Subpixel Phase Correlation Method Using Singular Value Decomposition and Unified Random Sample Consensus

Subpixel translation estimation using phase correlation is a fundamental task for numerous applications in the remote sensing community. The major drawback of the existing subpixel phase correlation methods lies in their sensitivity to corruption, including aliasing and noise, as well as the poor performance in the case of practical remote sensing data. This paper presents a novel subpixel phase correlation method using singular value decomposition (SVD) and the unified random sample consensus (RANSAC) algorithm. In the proposed method, SVD theoretically converts the translation estimation problem to one dimensions for simplicity and efficiency, and the unified RANSAC algorithm acts as a robust estimator for the line fitting, in this case for the high accuracy, stability, and robustness. The proposed method integrates the advantages of Hoge's method and the RANSAC algorithm and avoids the corresponding shortfalls of the original phase correlation method based only on SVD. A pixel-to-pixel dense matching scheme on the basis of the proposed method is also developed for practical image registration. Experiments with both simulated and real data were carried out to test the proposed method. In the simulated case, the comparative results estimated from the generated synthetic image pairs indicate that the proposed method outperforms the other existing methods in the presence of both aliasing and noise, in both accuracy and robustness. Moreover, the pixel locking effect that commonly occurs in subpixel matching was also investigated. The degree of pixel locking effect was found to be significantly weakened by the proposed method, as compared with the original Hoge's method. In the real data case, experiments using different bands of ZY-3 multispectral sensor-corrected images demonstrate the promising performance and feasibility of the proposed method, which is able to identify seams of the image stitching between sub-charge-coupled device units.

Research of Algorithm and Localization of Robot Based on Image Registration Technology

In order to improve the localization capability of the robot, digital image matching which based on the frequency domain was adopted, the method could reduce the impact of environmental factors, and adopts the sub-pixel subdivision pixel matching technology through curve fitting, Experimental results show that this method can achieve accurate localization in manipulator assembly working .

Small Baseline Stereo Matching Method based on Maximum Likelihood Estimation

A small baseline stereo matching method based on maximum like estimation is proposed. Firstly, the mixed window strategy is used to determine the matching window for each point in the reference image. Secondly, the normalized cross correlation function is employed to compute matching cost for possible matching points in the disparity range, then the “winner-take all” strategy is utilized to compute the disparity. Finally, the sub-pixel stereo matching method based on maximum like estimation is developed to obtain the sub-pixel disparity. The experimental results show that this method efficiently decreases adhesion phenomenon and achieves more accurate sub-pixel disparity

A FAST AND ROBUST SCENE MATCHING METHOD FOR NAVIGATION

The selection of matching method is critical to the scene matching navigation system, as it determines the accuracy of navigation. A coarse-to-flne matching method, which combines the area-based and feature-based matching method, is presented to meet the requirements of navigation, including the real-time performance, the sub-pixel accuracy and the robustness. In the coarse matching stage, the real-time performance is achieved by a pyramid multi-resolution technique, and the robustness is improved by multi-scale circular template fusion. In the precise matching stage, an improved SIFT method is introduced to calculate the matching position and the rotation angle. To validate the method, some experiments are completed. The results show that the proposed method can achieve the sub-pixel matching accuracy and improve the angle accuracy to 0.1 - .

Sub-Pixel Image Matching Based on the Fractal Dimension

The technology of image matching is very important in the field of image analysis. Studying on image correlation is the key to improve the accuracy of image matching, while the fractal self-similarity has the unique potential in making full use of image correlation. This paper put forwards a matching algorithm of the sub-pixel image matching based on fractal dimension. Firstly, by searching template subarea based on traverse in being matched image, and the sub-subareas are classified. Then, the fractal dimension of subareas and sub-subareas is calculated by using the optimal box counting method. The similarity maximum position can be found according to the statistical correlation. Finally, fractal interpolation method is used to improve the precision of image matching. Both theoretical analysis and experimental results show that this approach has good matching effect.

Subpixel image matching based on Fourier phase correlation for Radarsat-2 stereo-radargrammetry

Image matching is the major step in the radargrammetric process to measure elevation parallax. To extract parallax from stereo synthetic aperture radar images the subpixel image matching method based on Fourier phase correlation was implemented with an algorithm using the hierarchical multiresolution approach and applied to Fine Quad mode Radarsat-2 data. The experimental results with simulated images show that a decrease in intersection angle leads to an increase in matching accuracy of up to 0.06 of a pixel. To validate the matching results a digital surface model was extracted from the real stereo pair and compared with accurate lidar data. The statistics show that there are good improvements (in the order of 10%–20%) in the accuracy over results extracted using a traditional image matching technique based on the normalized cross-correlation. The analysis of the mutual dependence of matching accuracy and stereo pair configurations shows that the application of subpixel matching allows us to make the radargrammetric process flexible in the choice of stereo pairs.

IMPROVEMENT OF MEASUREMENT ACCURACY OF STRAIN OF THIN FILM BY CCD CAMERA WITH A TEMPLATE MATCHING METHOD USING THE 2ND-ORDER POLYNOMIAL INTERPOLATION

In this study, a system for non-contact in-situ measurement of strain during tensile test of thin films by using CCD camera with marking surface of specimen by black pen was implemented as a sensing device. To improve accuracy of measurement when CCD camera is used, this paper proposed a new method for measuring strain during tensile test of specimen with micrometer size. The size of pixel of CCD camera determines resolution of measurement, but the size of pixel can not satisfy the resolution required in tensile test of thin film because the extension of the specimen is very small during the tensile test. To increase resolution of measurement, the suggested method performs an accurate subpixel matching by applying 2nd order polynomial interpolation method to the conventional template matching. The algorithm was developed to calculate location of subpixel providing the best matching value by performing single dimensional polynomial interpolation from the results of pixel-based matching at a local region of image. The measurement resolution was less than 0.01 times of original pixel size. To verify the reliability of the system, the tensile test for the BeNi thin film was performed, which is widely used as a material in micro-probe tip. Tensile tests were performed and strains were measured using the proposed method and also the capacitance type displacement sensor for comparison. It is demonstrated that the new strain measurement system can effectively describe a behavior of materials after yield during the tensile test of the specimen at microscale with easy setup and better accuracy.

Efficient Path-Based Stereo Matching With Subpixel Accuracy

This paper presents an efficient algorithm to achieve accurate subpixel matchings for calculating correspondences between stereo images based on a path-based matching algorithm. Compared with point-by-point stereo-matching algorithms, path-based algorithms resolve local ambiguities by maximizing the cross correlation (or other measurements) along a path, which can be implemented efficiently using dynamic programming. An effect of the global matching criterion is that cross correlations at all pixels contribute to the criterion; since cross correlation can change significantly even with subpixel changes, to achieve subpixel accuracy, it is no longer sufficient to first find the path that maximizes the criterion at integer pixel locations and then refine to subpixel accuracy. In this paper, by writing bilinear interpolation using integral images, we show that cross correlations at all subpixel locations can be computed efficiently and, thus, lead to a subpixel accuracy path-based matching algorithm. Our results show the feasibility of the method and illustrate significant improvement over existing path-based matching methods.

Automatic Deformation Detection for Aircraft Engine Disk Inspection

Computer vision algorithms are seeing increased use in industrial inspection applications. Here, we present an “Aid to Visual” system that can detect post deformations of less than 0.005 inches in jet engine high pressure turbine disks. We create a gold-standard reference post from the posts of sample turbine disks and then use registration, edge detection, and curve-similarity algorithms to identify unacceptable post deformations. We address the challenges associated with adapting academic algorithms for use in functioning inspection systems. We present novel solutions to deal with practical issues such as accuracy, speed, robustness, and ease of use. We also present a novel, highly-efficient sub-pixel contour matching algorithm and demonstrate the effectiveness of using sub-pixel distance calculation. We demonstrate overall error rates less than 1% on over 2400 images of posts. We have integrated our algorithms into the commercial LabVIEW software running on the Aid To Visual workstation. Our algorithms will enable plant-factory inspectors to identify minute post deformations that were previously difficult to detect.

Auto-adjusted 3-D optic disk viewing from low-resolution stereo fundus image

Three-dimensional (3-D) visualization of the optic nerve head (optic disk) is very useful for clinical applications. It allows clinicians to measure the disk parameters more accurately and thus make the pathological diagnosis and progression monitoring easier. This paper describes an automatic, precise, 3-D optic nerve head reconstruction method from a pair of stereo images for which efficient steps including sparse-image registration and dense-depth recovery are used. A combination of two registration methods is used to detect the sub-pixel correspondences. The proposed method takes advantages of both the correlation methods which is robust to noise and the feature-based method on its accuracy. The searching range in image registration is auto-adjusted based on the previous iteration result. Only sparse matched points are computed to speed up the processing and the sub-pixel matching is used to overcome the problem of low resolution in the image. This is followed by the piecewise cubic interpolation to obtain the dense disparities and depths. Multiple windowing is applied here by first using the large window to obtain basic disparities followed by the small window and previous basic disparities to measure details. The result is then smoothed and displayed as the final 3-D shape.

Point matching under large image deformations and illumination changes

To solve the general point correspondence problem in which the underlying transformation between image patches is represented by a homography, a solution based on extensive use of first order differential techniques is proposed. We integrate in a single robust M-estimation framework the traditional optical flow method and matching of local color distributions. These distributions are computed with spatially oriented kernels in the 5D joint spatial/color space. The estimation process is initiated at the third level of a Gaussian pyramid, uses only local information, and the illumination changes between the two images are also taken into account. Subpixel matching accuracy is achieved under large projective distortions significantly exceeding the performance of any of the two components alone. As an application, the correspondence algorithm is employed in oriented tracking of objects.

High-Precision Range Estimation from an Omnidirectional Stereo System.

This paper presents the estimation of depth based on the cylindrical re-projection of images captured by a stereo setup of ODV (OmniDirectional Vision). In general, images obtained by such a device have relatively low spatial resolution and have some aberration in comparison to standard camera images and thus, considerable range estimation errors are caused. We propose a stereo setup of ODV, which features two mirrors with the original curvature to minimize blurring influence. We also propose an edge-based intensity interpolation method to improve the spatial resolution of the developed image. The bidirectional subpixel matching procedure and the discontinuity compensation algorithm are also presented to improve the quality of the range estimation. The experimental result reveals that the proposed ODSV (OmniDirectional Stereo Vision system) can successfully contribute to reduce disparity error and has subpixel range estimation accuracy even in the cluttered outdoor environment.

Quantitative 3-d diagnostic ultrasound imaging using a modified transducer array and an automated image tracking technique.

An approach for acquiring dimensionally accurate three-dimensional (3-D) ultrasound data from multiple 2-D image planes is presented. This is based on the use of a modified linear-phased array comprising a central imaging array that acquires multiple, essentially parallel, 2-D slices as the transducer is translated over the tissue of interest. Small, perpendicularly oriented, tracking arrays are integrally mounted on each end of the imaging transducer. As the transducer is translated in an elevational direction with respect to the central imaging array, the images obtained by the tracking arrays remain largely coplanar. The motion between successive tracking images is determined using a minimum sum of absolute difference (MSAD) image matching technique with subpixel matching resolution. An initial phantom scanning-based test of a prototype 8 MHz array indicates that linear dimensional accuracy of 4.6% (2 sigma) is achievable. This result compares favorably with those obtained using an assumed average velocity [31.5% (2 sigma) accuracy] and using an approach based on measuring image-to-image decorrelation [8.4% (2 sigma) accuracy]. The prototype array and imaging system were also tested in a clinical environment, and early results suggest that the approach has the potential to enable a low cost, rapid, screening method for detecting carotid artery stenosis. The average time for performing a screening test for carotid stenosis was reduced from an average of 45 minutes using 2-D duplex Doppler to 12 minutes using the new 3-D scanning approach.

Direct linear sub-pixel correlation by incorporation of neighbor pixels' information and robust estimation of window transformation

Standard methods for sub-pixel matching are iterative and nonlinear; they are also sensitive to false initialization and window deformation. In this paper, we present a linear method that incorporates information from neighboring pixels. Two algorithms are presented: one ‘fast’ and one ‘robust’. They both start from an initial rough estimate of the matching. The fast one is suitable for pairs of images requiring negligible window deformation. The robust method is slower but more general and more precise. It eliminates false matches in the initialization by using robust estimation of the local affine deformation. The first algorithm attains an accuracy of 0.05 pixels for interest points and 0.06 for random points in the translational case. For the general case, if the deformation is small, the second method gives an accuracy of 0.05 pixels; while for large deformation, it gives an accuracy of about 0.06 pixels for points of interest and 0.10 pixels for random points. They are very few false matches in all cases, even if there are many in the initialization.