Open AccessPanoramic Image Stitching with Efficient Brightness Fusion Using RANSAC Algorithm
Author(s) -
Jungpil Shin,
Abdur Rahim,
Keun Soo Yun
Publication year - 2018
Publication title -
international journal of engineering and technology
Language(s) - English
Resource type - Journals
ISSN - 2227-524X
DOI - 10.14419/ijet.v7i3.34.18981
Subject(s) - image stitching , ransac , artificial intelligence , computer vision , brightness , hamming distance , computer science , matching (statistics) , image fusion , binary number , homography , image (mathematics) , pattern recognition (psychology) , mathematics , algorithm , optics , statistics , physics , arithmetic , projective test , projective space
Background/Objectives: Image stitching can enhance the picture very pleasant by modifying and mixing the different aspects.Therefore, we present panoramic image stitching with efficient brightness fusion which is challenging in different bright sequences taken from different angles.Methods/Statistical analysis:For the problem of brightness, the input image is mixed with sequential images in different brightness.In this works, we proposed atechnique that blends multiple brightness using simple quality measures like color, saturation and contrast.The resulting image quality is good, and most important thing is, the method is efficient since it is simple. Thanthe resulting fused images is applied for panorama image stitching. We used multiband blending to prevent the blurring, and BRIEF (binary robust independent elementary features) method for feature descriptors. We solved the multi-image matching problem using Hamming Distanceusing the binary string based descriptors which is most similar features compare with the second most similar images.We proposed FLANN based matcher to get the more accurate results for using large datasets. We estimate Homography with the matching images using RANSAC algorithms.Findings:An effective structure is performed when we are able to resolve the brightness correction in expose too much or expose for too short a time and the appearance ghost. To solve the unification of brightness, we have collected the input images in different exposures, and selection of the good parts of each picture to an input image for stitching.We removed the blurring from input images, and solved multi-image matching using Hamming distance method. We found better results comparing other methods. For large dataset, we used FLANN based matcher, and estimated the Homography using RANSAC algorithm.Improvements/Applications:We have shown the performance of panoramic image stitching with efficient brightness fusion. We performed stitching with high regulationimages. Finally, we were able to create a panoramic image with efficient brightness fusion.