An Algorithm to Detect Color Based Objects with its Current Moving Direction for UAVs

Abstract

Human concentration is required in typical surveillance system to monitor behaviour of object in complex motion. To relieve human interaction and to process risk detection and analysis automatic vision based detection system is required. The object of interest which is moving in frames of video at each moment can be tracked by constantly assigning tags to object. Aerial robots are mostly used in surveillance, rescue operation and where ground navigation is difficult. Quad-copter has advantage of higher payload due to its four rotors and it makes it more suitable to carry transmitter equipment and camera.   In this paper, an algorithm for detection and tracking of object based on color features in real time is presented for the unmanned aerial vehicles (UAVs) . The proposed algorithm can detect moving objects in real time, track the object with contrail and can also provide direction of the object in which it is moving.  The object is detected and its position N(x.y) is registered in the list and with the movement the contrail is formed. By comparing the position of object in the frame, the direction of movement is given. The algorithm worked efficiently in real time.