KENDALI A.R DRONE BERDASARKAN CITRA TANGAN MENGGUNAKAN METODE HAAR CLASSIFIER; A.R DRONE CONTROL BASED ON HAND IMAGE USING HAAR CLASSIFIER METHOD

Adisatrio, Bimo Adisatrio F.A, Bakhtiar Alldino

2016 | Skripsi | FMIPA

Haar Cascade Classifier Algorithm is a method founded by Viola – Jones and has been proved robust to detect object. Implementation of this method in AR.Drone becomes human-machine interface development. In this research, a prototype system has been built to detect hand position and control the drone by identify hand position of operator. Digital image processing has been implemented on the system by using library opencv with Haar Classifier method. Frame input from AR.Drone`s camera sent to computer for detection processing and classify the centroid. Results from this research is a system that is able to recognize the position of the operator's hand automatically using an XML file system data on the training results. Trials done by testing operator distance to the AR.Drone and AR.Drone respond to commands with parameter variate detection. The best result is when the operator is standing upright and facing the camera AR.Drone in demonstrating the hand position frontally. The result proven that AR.Drone can perform the mission by detecting hand position. AR.Drone can detect hand position in range within 2 - 2.8 metres. System using an XML file with number of positive image 768 with an average response time of detection is of 0.05 seconds.

Kata Kunci : position, hand signal, drone.