IMPLEMENTASI METODE HOUGH TRANSFORM UNTUK VISUAL TRACKING JALAN RAYA DARI UDARA DENGAN MEMANFAATKAN UAV; IMPLEMENTATION OF HOUGH TRANSFORM METHOD FOR VISUAL HIGHWAYS TRACKING USING UAV

Santoso, Dimas A, R.Sumiharto

2015 | Skripsi | FMIPA

This study has been carried out the implementation of a visual tracking system highways using Unmanned Aerial Vehicle (UAV) using a Single Board Computer (SBC) as image processing. Implementation of image processing consists of image preprocessing process, the detection process highways, and highway tracking process that is processed in the device SBC Cubieboard type 2 (A20). Data input to the image processing in the form of a video image captured highways and processed in real time when the aircraft was flown. The method used is that image segmentation system with HSV thresholding, Canny and Hough Transform for the detection process highways and calculating the average value of the coordinates of the midpoint of the highway. As well as methods for tracking Moments highway. The process of detection and tracking of these roads is done by image processing using OpenCV library with programming language C ++. System testing begins with the calibration of color values of the highway to look for HSV as a reference value detection process highways. Then proceed with testing visual tracking system at three different locations namely Imogiri Road East, West Ringroad Roads and Street South Ringroad. Highway used asphalt manifold and has a surrounding environment that color contrast with the road or in the form of vegetation. From the test results tracking highway, Moments methods used can provide the best results if the result of good color segmentation highway, combined with the calculation of average lines resulting from the Hough Transform method. The final result of the visual system is the coordinate position tracking and video highway that has been processed through thresholding process, dilation, Canny, Hough Transform, ROI, and the calculation method of Moments.

Kata Kunci : Cubieboard; real time; UAV; Moments; and OpenCV