Sledování objektů zájmu pomocí kvadkoptéry

Abstract

This work deals with detections and tracking objects of interest (people), processing the image obtained from the camera mounted on quadcopter. SVM detector and HOG descriptor are used for the target detection, both implemented in OpenCV library. These detections are inserted into the buffer and processed by the algorithm selecting the key detection. Theoretical part is focused on the drone description and the computer vision in general. Practical part describes the program itself, its classes, algorithms or solutions. Source codes are written in object orientated scripting programming language Python.