Computer vision for drone control: development by a student of SSUGT
Anastasia Filkova, a student at the Siberian State University of Geosystems and Technologies, has developed an autonomous control system based on computer vision. Drones are now widely used in various fields, from logistics to agriculture and rescue operations. However, their mass implementation is hampered by the need for constant operator control.
The system turns drones into «smart» assistants that can independently track objects, regulate distances, and avoid obstacles. It is based on a framework based on OpenCV and Pioneer SDK using Python. To improve efficiency, the NumPy (mathematical calculations) and PyQt5 (graphical interface) libraries are used.
Key features include automatic detection and tracking of objects in real time (15+ frames per second), adaptive flight control with dynamic adjustment of parameters, as well as energy-saving modes to increase operating time.
The system has undergone testing at the SSUGT artificial intelligence laboratory and confirmed the expected results.
This development opens up new prospects for commercial and industrial use of drones – from monitoring construction sites and filming dynamic events to search and rescue operations.
