IMPLEMENTASI METODE HAAR UNTUK MELACAK WAJAH SECARA OTOMATIS PADA QUADROTOR; HAAR METHOD IMPLEMENTATION FOR FACE TRACKING AUTOMATICALLY ON QUADROTOR

Noto, Fakih Arief, N

2015 | Skripsi | FMIPA

Utilization of quadrotor to take photo is being trend specially for aerial photo and video lovers. Photo and video object involve human as object should not lose focus to the human’s face to get good result. This research planned and built a system for face tracking on quadrotor which can follow face moving automatically. Image processing is implemented on computer use wireless communication between computer and quadrotor to get image. Library used are Ar Drone .Net OpenCV, and EmguCV Library with Haar cascade classifier for face detection. The result of face detection used for controlling quadrotor to keep face presentation at the centre of image and keep distance between quadrotor and face. The results of this research is system in quadrotor which can move quadrotor for face tracking with haar cascade classifier face detection method. The results of testing are that environtment with light intensity between 3,10 cd (39 lux) until 195,76 cd (2460 lux) are best for face detection. The best angle of facial position shift for face detection is below 20°. Declivity roll angle’s quadrotor for face detection is from -18° until 19°. Controling priority from pitch move, roll move and throttle. Maximal velocity for backward forward move = 40 rectangle width/second, axisX image move = 20 pixel/second, axisY image move = 42 pixel/second.

Kata Kunci : cascade classifier; EmguCV; Ar Drone quadrotor; OpenCV