Accelerated RANSAC for Accurate Image Registration in Aerial Video Surveillance

Abstract

Compared with ground views and direct overhead views (for orbital satellites), aerial robotics allow for capturing videos from diverse viewpoints and scenes, thus, the content of aerial image is complex and changeable, and aerial video has complex inter-frame transforms stemming from the blend of camera motion, platform motion and jitter. In addition, poor quality and similar texture are common in longdistance and large-scale aerial video surveillance. All of these interferences make image registration of aerial video difficult. This paper puts forward one image registration method suited to aerial video on the basis of the hypothesize-and-verify of RANSAC. The proposed accelerated RANSAC, named PSSC-RANSAC (Prior Sampling & Sample Check RANSAC), incorporates prior sampling, which comes from three levels of sample evaluation, including texture magnitude, spatial consistency and feature similarity, to generate more possibly correct samples in priority. Furthermore, prior information of sample, quality of sample subset and subset invariability are together used to check the sample subsets, and the incompatible arrangements of subsets are immediately ruled out in sample check stage, which speeds up the iteration further. Results of the experiment have proved the good performance of the presented PSSC-RANSAC at 90% contamination level. For typical image pairs, the number of iterations is reduced by at least 16.67% and evaluation computation is reduced by at least 11.01% compared with SVH-RANSAC, and the reprojection error is decreased by at least 4.44% and 6.31% compared with RANSAC and SVH-RANSAC, respectively. It can overcome the interferences, and is very suitable for image registration of aerial images.

RANSAC for Robotic Applications: A Survey

“…In this case, SURF is the method used to detect features, and the authors find that it compares favorably to using SURF, SIFT, or ORB alone, although their experimental setup only includes a pair of aerial images. In [ 75 ], the authors present Prior Sampling and Sample Check RANSAC (PSSC-RANSAC), which incorporates prior knowledge of the sampling goodness coming from three different sources: texture magnitude, spatial consistency, and feature similarity. This prior sampling should possibly generate more correct samples.…”

Image matching algorithm based on improved AKAZE and Gaussian mixture model

“…16 In the feature matching step, the descriptors obtained in the previous step are matched using similarity or distance measures, and the most popular algorithms are brute-force (BF) 17 matcher, fast library for approximate nearest neighbors, 18 and random sample consensus (RANSAC). 19 Image matching seeks to identify and align content or structures with the same or similar features in two photos at the pixel level. The commonly used algorithms for image matching are divided into two categories: region-based image matching methods 20 and featurebased image matching methods.…”

“…The main feature descriptor algorithms currently in use are BRIEF, 14 histogram of oriented gradient, 15 and local binary pattern 16 . In the feature matching step, the descriptors obtained in the previous step are matched using similarity or distance measures, and the most popular algorithms are brute-force (BF) 17 matcher, fast library for approximate nearest neighbors, 18 and random sample consensus (RANSAC) 19 . Image matching seeks to identify and align content or structures with the same or similar features in two photos at the pixel level.…”

Multi-Angle Optical Image Automatic Registration by Combining Point and Line Features

“…Zheng put forward an image registration method based on RANSAC (Random Sample Consensus), which is suitable for processing aerial video. It incorporates prior sampling to possibly generate more correct samples [9]. These methods have their own characteristics, but they are not designed for multi-angle images.…”