SISTEM PENTAUTAN FOTO UDARA MENGGUNAKAN DETEKSI FITUR DENGAN ALGORITMA HARRIS CORNER DETECTION (HCD); AERIAL IMAGE STITCHING SYSTEM BASED ON HARRIS CORNER DETECTION (HCD) FEATURE DETECTION ALGORITHM

Aniq, Nilta, R. Sumiharto

2015 | Skripsi | FMIPA

Remote sensing is the science of obtaining information about an object, area, or symptoms with the analyze data acquired using the tool without contacts directly towards the object. aerial image is one way to remote sensing. On the mapping of an area, it takes a lot of aerial images with broad coverage with a clear sight of the object so that it is easier in the analysis of the data. Thus, in this research, aerial image stitching system based on HCD feature detection is selected that can generate the image with a wider coverage and the clear object This system is designed to be able to merge the aerial image using HCD, KDtree, RANSAC and warpPerspective algorithm. HCD algorithm is used for the detection of point features in the form of a corner of each image input, the algorithm kdtree as feature matcher, ransac algorithm to reduce outliers and produce homography matrix that used on warping process with the warpperspective algorithm. This research also use KNN as matcher to find out the influence of the matcher against merger with the same detection method. System performance tested with a wide variety of rotation, scale, and translation. KDTree performed well when tested with variety of rotation, with images that cannot merge well have a rotation 110o, 140o, 300o and 320o. When system tested with variety of scale, KNN performed better than KDTree, with the result is system can merge the images that have scale 80%, 90%, 100%, 110% and 120%, and when the system tested with variety of translation, KNN performed better, with the result is system can merge images with all translation, except image with translation 10% horizontal and 10% vertical.

Kata Kunci : aerial image; HCD; KDTree; KNN; RANSAC; warpperspective