Vehicle window recognition system based on projection and hough straight line detection

Abstract

The present invention discloses a vehicle window recognition system based on projection and hough straight line detection. The system comprises: a vehicle window region coarse positioning unit, a vehicle window upper and lower border positioning unit and a vehicle window left and right border positioning unit which are sequentially connected; the vehicle window region coarse positioning unit primarily positions a vehicle window region based on detection of a vehicle license plate position, and in the coarsely positioned vehicle window region, the vehicle window upper and lower border positioning unit, via a hough straight line detection module and a horizontal gradient processing module and by using technologies of hough straight line detection, horizontal gradient and the like, respectively performs processing on information of the coarsely positioned vehicle window region so as to find optimum upper and lower border positions of a vehicle window and determine upper and lower borders of the vehicle window; and via the vehicle window left and right border positioning unit and by using the hough straight line detection technology, suitable left and right border positions of the vehicle window are found and left and right borders of the vehicle window are determined so as to detect the accurate vehicle window region and complete the accurate recognition detection of the vehicle window.

Description

technical field [0001] The invention relates to the field of vehicle identification, in particular to a vehicle window identification system based on projection and hough line detection. Background technique [0002] In recent decades, intelligent transportation systems have achieved rapid development. Vehicle window area recognition is a relatively new research direction in intelligent transportation systems, and vehicle window detection is an important function in bayonet vehicle recognition systems. When the window area is detected, a series of vehicle fingerprint detection such as annual inspection marks, co-pilot personnel, and front decorations can be carried out on the basis of it, which provides a basis for the vehicle identification system and improves the performance of vehicle identification to a large extent. Accuracy, therefore, the detection of car windows is a key technology in the vehicle recognition system. [0003] The current window detection algorithm m...

AI Technical Summary

Problems solved by technology

These two methods have the following disadvantages: when using the gradient-based window detection method, if the window boundary is not obvious enough, the gradient of its border is not obvious; and when the window detection method based on color difference, if the window area The color is not sufficiently different from the body, and the window detection will not be very accurate

Images