| IMPROVE THE MATCHING CAPABILITY OF SURF ALGORITHM BY PRE-PROCESSING OF IMAGES |
| Paper ID : 1025-SMPR |
| Authors: |
|
Seyed Komeil Mahmoudi, Mohammad Sedighi *, Amir Shahrokh Amini Dept. of Geomatic Engineering, Islamic Azad University South Tehran Branch, Piroozi Street, Abuzar Boulevard, Dehhaghi Avenue, P. O. Box: 1777613651, Tehran, Iran |
| Abstract: |
| Feature-matching algorithms, such as the SURF1 algorithm, have been very successful in recent years. Nowadays, with the development of computer hardware, operations such as image matching are performed with high speed, although they involve comparing a large number of key points together. The SURF matching algorithm is appropriate method in terms of scale and time and has high stability against image noise and brightness variations. The purpose of this paper is to improve the capabilities of the SURF algorithm in the matching procedure by adding pre-processing operations to the algorithm. In the pre-processing stage, image enhancement is performed, which includes all the processes that extract the best and most complete information in the image such as using the Laplacian filter. The proposed method introduces a series of Laplacian sharpening filters to detect more feature points. The results of investigation showed that using the proposed pre-processing operation stage improves the results of image matching. |
| Keywords: |
| Matching, Remote Sensing, Pre-processing, Image Enhancement, Surf Algorithm |
| Status : Conditional Accept (Poster) |
