Vehicle classification has evolved into a significant subject of study due to its importance in autonomous navigation, traffic analysis, surveillance and security systems, and transportation management. While numerous approaches have been introduced for this purpose, no specific study has been conducted to provide a robust and complete video-based vehicle classification system based on the rear-side view where the camera's field of view is directly behind the vehicle.
Prof. Bhanu and his colleagues from the University of California, Riverside have developed a method for recognizing the logo of a vehicle from a low-resolution video feed in real-time. This method works by using a software system for super-resolving the vehicle maker's logos, which facilitates recognition of a vehicle make more reliably than low-resolution vehicle logos. A super-resolution algorithm produces a high-resolution image from low-resolution video input. This eliminates the need for high resolution and close up images of the logo, license plate, and rear view by using simple low-level features like the vehicle width and height. As a result, this method is computationally inexpensive and the method is capable of running in real-time.
Figure 1:Shadow removal and obtaining the bounding box for vehicle identification
|United States Of America||Issued Patent||10,127,437||11/13/2018||2012-885|