SISTEM PENGGABUNGAN CITRA UAV DENGAN MENGGUNAKAN METODE BRISK (BINARY ROBUST INVARIANT SCALABLE KEYPOINTS); UAV IMAGE STITCHING SYSTEM BASED ON BRISK (BINARY ROBUST INVARIANT SCALABLE KEYPOINTS) METHOD

Hapsari; Noviaris, R. Sumiharto

2015 | Skripsi | FMIPA

Image data taken by the UAV are used as mapping, the captured image is obtained not cover the whole territory of the desired object. This can be solved by using the technique of stitching image. The problems that occur in the application of image stitching is the change in the rotation and scale in the image so result that the image stitching that many of which can not be combined. In this research applied a method BRISK to increase resistance to changing the image of the rotation and scale. BRISK method is used as detection and extraction in the process stitching image. This method has a process of making space scale aimed to detection of keypoint so that this process makes the method BRISK resistant to change scale. In addition, there are estimates of the keypoint orientation and rotation patterns of the sample so that the method BRISK be resistant to rotation. In this system has other stages that keypoint matching and merging images. Keypoint matching aims to search for the same keypoint between image 1 to image 2 by using the method of FLANN. Then merging the image of previously conducted search homography matrix in each image by using RANSAC and combined using warpPerspective. Based on the results of testing known that the variety of rotation that can be combined from the angle of 0° to 360°, the minimum number keypoint which corresponds to acquire the image of the mosaic that is 4, the scale of the maximum obtained from testing variations of the scale is 460%, minimum scale obtained from testing variations scale is 33,3%, and the results of testing a minimum of overlap that can be combined is 10% for the horizontal axis and 11.3% for the vertical axis.

Kata Kunci : Image Stitching; Aerial Photography; BRISK; FLANN; RANSAC