TY - GEN
T1 - Image stitching development by combining SIFT Detector and SURF descriptor for aerial view images
AU - Akhyar, Ramaulvi Muhammad
AU - Tjandrasa, Handayani
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/7
Y1 - 2019/7
N2 - Image stitching is considered as a subject that can be researched in the field of computer vision. Aerial view images are the type of images that requires image stitching because aerial images can be taken based on the flying path of a drone so that the images obtained are in the form of sequences, and the research that examines the use of different types of feature detector and descriptor in aerial view images is still lacking, therefore this research will combine two different types of feature detector and descriptor to get a better image stitching result. In the previous study the SIFT-SURF combination was used as image matching and succeeded by obtaining a small time cost and high accuracy, because of that in this study SIFT is used as a feature detector and SURF is used as a feature descriptor with input data of aerial view images. Features matching is carried out by applying a bruteforce matcher using an euclidean distance, after the initial matches or raw matches are obtained, the Lowe ratio test is done to eliminate false matches, after that Random Sample Consensus (RANSAC) is utilized as homography which is used to get correct matching that can be called inliners. The last step is image warping to adjust the obtained matches. The measurement used in this study is the number of obtained features, the number of good matches, and the time cost as the parameters to get the system performance. This research compares the proposed method with other combination methods, namely SURF-SIFT, and BRISK-ORB. This research proposed an image stitching system to stitch aerial view images by using 5 pairs of an aerial view images. The image stitching system is designed with the several steps which is preprocessing, SIFT detector and SURF description, euclidean distance matching, Lowe ratio test, RANSAC and images stitching. From the experiment that has been done, SIFT-SURF combination successfully stitch the tested images with less computational time and it also have more or the same good matching with less detected feature.
AB - Image stitching is considered as a subject that can be researched in the field of computer vision. Aerial view images are the type of images that requires image stitching because aerial images can be taken based on the flying path of a drone so that the images obtained are in the form of sequences, and the research that examines the use of different types of feature detector and descriptor in aerial view images is still lacking, therefore this research will combine two different types of feature detector and descriptor to get a better image stitching result. In the previous study the SIFT-SURF combination was used as image matching and succeeded by obtaining a small time cost and high accuracy, because of that in this study SIFT is used as a feature detector and SURF is used as a feature descriptor with input data of aerial view images. Features matching is carried out by applying a bruteforce matcher using an euclidean distance, after the initial matches or raw matches are obtained, the Lowe ratio test is done to eliminate false matches, after that Random Sample Consensus (RANSAC) is utilized as homography which is used to get correct matching that can be called inliners. The last step is image warping to adjust the obtained matches. The measurement used in this study is the number of obtained features, the number of good matches, and the time cost as the parameters to get the system performance. This research compares the proposed method with other combination methods, namely SURF-SIFT, and BRISK-ORB. This research proposed an image stitching system to stitch aerial view images by using 5 pairs of an aerial view images. The image stitching system is designed with the several steps which is preprocessing, SIFT detector and SURF description, euclidean distance matching, Lowe ratio test, RANSAC and images stitching. From the experiment that has been done, SIFT-SURF combination successfully stitch the tested images with less computational time and it also have more or the same good matching with less detected feature.
KW - Euclidean Distance
KW - Feature Descriptor
KW - Feature Detector
KW - Image Stitching
KW - Random Sample Consensus
KW - Scale Invariant Feature Transform
KW - Speeded Up Robust Feature
UR - http://www.scopus.com/inward/record.url?scp=85073536542&partnerID=8YFLogxK
U2 - 10.1109/ICTS.2019.8850941
DO - 10.1109/ICTS.2019.8850941
M3 - Conference contribution
AN - SCOPUS:85073536542
T3 - Proceedings of 2019 International Conference on Information and Communication Technology and Systems, ICTS 2019
SP - 209
EP - 214
BT - Proceedings of 2019 International Conference on Information and Communication Technology and Systems, ICTS 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 12th International Conference on Information and Communication Technology and Systems, ICTS 2019
Y2 - 18 July 2019
ER -