The Application of Spatio-Temporal Video Fusion Technology in Dynamic Target Trajectory Tracking in the Power Industry

Traditional video analysis or GPS single location technologies have been limited in complex electricity environments in recent years, which means some researchers have used more advanced analytical methods. A study in this thesis tested the practicality of using spatio-temporal video fusion technology to track dynamic target trajectory in the electricity industry. The study results were collected from experiments involving three different aerial paths that were carried out under three different environmental conditions: clear sky, overcast sky, and night. Different indicators are evaluated, trajectory extraction accuracy, spatio-temporal positioning accuracy, and video processing time. The results of the study reveal that in terms of the central indicator of trajectory extraction accuracy, it can hold its accuracy rate while under different atmosphere conditions, with the highest trajectory extraction accuracy of 99.7%. The video processing times were measured to be 1.2-1.8 seconds in clear sky conditions, 1.6-2.1 seconds in cloudy conditions, and 2-3.2 seconds in night conditions, and revealed that the system can be processed in a pretty quick time.