Most of the vehicle tracking methods based on images use visible images, but it is difficult for such methods to adjust to changes in the weather and to the different lighting conditions of day and night. However, vehicles must be tracked in all weathers and at any time. Therefore, in this study, the authors propose a method using infrared images that can be applied in all weathers and at any time. Because of the characteristics of infrared images, the project does not use template matching, and mainly applies abackground subtraction method for vehicle tracking.
Samenvatting