Vision-based autonomous object tracking using multi-rotor UAV

In this paper- we are presenting a real-time method to detect and track an autonomous objectusing visual processing on an unmanned aerial vehicle using an on-board companioncomputer (Jetson-TX1) for image processing. The profile of objects, frame rate of images,and unexpected motion make it hard to detect and track the object for a long period. To caterto this, we came up with an algorithm that was developed for long-term tracking which makesuse of Discriminative correlation filter with Channel and Spatial Reliability. The majorrestriction of our algorithm arises in the presence of occlusion, which was solved by creatinga region of interest in the center of the frame in which the object will always reside. If theobject exits the center region a command of left, right, top or bottom will be generatedfollowing the position of the object relative to our center position. These commands will becommunicated to the UAV via Mavlink protocol. Experimental results show that we haveachieved a long period of tracking with a good frame rate and eliminating spurious events andmisdetections.