THE PERFORMANCE EVALUATION OF ELECTRONIC VISUAL TRACKING ALGORITHMS WITH DIFFERENT SENSORS

The purpose of this paper is to evaluate the precision and activity of the object tracking algorithms ("trackers") as a relation function of camera resolution. To do this, five sensors (webcams) that have different resolutions are used to acquire the video sequences that cover most challenging states in object tracking. Five open-source visual trackers that are got as the assistances branch of the "Open Computer Vision (OpenCV)" collection, with the same initial state of the target, are run at those videos from which the trackers have to follow the moving target in the subsequence frames. The experiments were carried out using numerous estimation conditions to identify in what way that trackers execute. The quantitative and qualitative results were then analysed to recognize effective methods for robust tracking. The results confirmed that the tracker's speed performance was changed based on the visual sensor resolution. They also showed that the trackers' accuracy increased as the visual sensor resolution increased.