Detecting and tracking of traditional instrument with different camera vision

Abstract

In this paper, we present an approach for detecting, recognizing and tracking varying number of objects through both have similarity shape and colors. The objects tracking is the varying notes of Indonesian traditional instrument called “gambang”. Our method build an idea how to recognize well the notes and classify each objects in real time tracking. The houghlines standard method is used to search the right lines of the shape of instrument. At the level of detecting object, the square and contours method is used. The real time tracking is based on the color of contour object using hsv color filter. We demonstrate the viability of this approach by experimenting on several videos of the instrument from different angle of view 2D camera. The instrument labeling testing with the camera view from the left side has 100% success, from the center side has 95% success, and from the right side has 90% success.