Video System With Automatic Tracking of the Object of Interest

The article presented a comparison of SIFT, AKAZE, BRISK, ORB methods of searching feature points and CSRT, KCF, GOTURN methods of tracking by the criterion of speed (the number of processed frames per second), as well as by the number of feature points found from the specified error value. The software for tracking the position of the object of interest in the video stream was developed based on these methods and using the Python programming language. A graphical user interface (GUI) for the software for tracking the object of interest was created using the Qt Designer software. A system consisting of a camera, a camera bracket, servos controlling the camera position, a microcontroller and a personal computer was built. A program for STM32F1 microcontrollers responsible for changing the camera position by servos was written using C++ programming language. The dependences of the camera rotation angles on the time when tracking the object were constructed. The trajectory of the tracked object was also constructed. Various data can be obtained from this trajectory, for example, you can find out the direction of movement, calculate speed and acceleration.