Rancang Bangun Camera Face Tracker dengan Menggunakan Metode Haar-Like Feature dan PID
Computer Vision is one branch of the Field Science Image Processing which allows computers to see like humans, recognize objects and make decisions one example application is face detection, which later developed as the face tracker. Haar-like feature is a feature based on the Haar wavelet. Haar wavelet is a single square wave (high and low). For two dimensions, one light and one dark. The workings of haar-like features in a face detection technique is to use a sliding window, the whole picture and try to see if there is a part of the image that is shaped like a face or not. Controller Proportional-Integral-Derivative (PID) closed loop control model is a widely used in industrial. This control model will calculate the error as the value of the difference between the measured results with the desired set point value. Controller PID will minimize the error by adjusting the process control inputs. In this study utilizing Haar-like feature and PID to perform the face tracker system, starting with the webcam detects a face, using EmguCV will determine the required set-point motor driver to drive the motor with a PID controller, thus motor will follow the movement of the face to the left and or right.
Based on the test results in this study, the system can recognize the faces with optimal response when
the value of Kp = 0.8, Ki = 0001 and Kd = 0.05 and an average error of 2%.