IMPLEMENTASI DETEKSI OBJEK BERBASIS WARNA PADA ROBOT SEPAK BOLA BERODA
GUMELAR, MUHAMMAD FAJAR
MetadataShow full item record
In the Indonesian Wheeled Soccer Robot Contest there are many regulations about the color of objects, such as the color of the ball, the goal, the mark of the opponent’s color, and the match field. Based on these factors, this research aims to build a color-based object detection system to detect those object. The detection process starts with transforming RGB into HSV, then opening, closing, then filters 2D Convolution, after that thresholding, and finally getting the color’s contour of the object, so moment analysis can be done and coordinates of the object’s center of gravity can be gained. The test results showed that Mr. Dev’s color-based object detection system is able to obtain data of ball’s coordinates and angles, shooting target’s coordinates and angles, enemy color marker’s coordinates and angles, and field’s edges data. Besides that, the ball color’s anti-noise system has also run well. The system is able to get an average FPS value of 25,324 on the front camera, and 20,174 on the omnivision camera.