Non-motor vehicle license plate area positioning method based on multi-feature fusion

Abstract

The invention discloses a non-motor vehicle license plate area positioning method based on multi-feature fusion. The method comprises the following steps: preprocessing an acquired to-be-positioned image; performing coarse positioning on a non-motor vehicle license plate area by adopting a color template matching algorithm based on an HSV color space; adopting a Sobel operator-based vertical edge detection algorithm to realize accurate positioning of a primary non-motor vehicle license plate area; adopting a character feature detection algorithm based on MSER features to carry out secondary accurate positioning on a non-motor vehicle license plate area; utilizing an SVM algorithm to judge the accurately positioned area, and removing non-license plate areas; and correcting the distorted non-motor vehicle license plate image by adopting an affine transformation method to realize accurate positioning of the non-motor vehicle license plate area. The non-motor vehicle license plate area positioning method is high in adaptability to complex environments and high in operation speed, can meet the accuracy and real-time performance requirements of non-motor vehicle license plate recognition, and provides high-quality non-motor vehicle license plate images for subsequent non-motor vehicle license plate character segmentation and recognition.

Description

technical field [0001] The invention relates to the fields of computer image processing and intelligent transportation, in particular to a multi-feature fusion non-motor vehicle license plate area positioning method. Background technique [0002] With the continuous improvement of people's living standards, non-motorized vehicles such as electric vehicles and bicycles have become an indispensable means of transportation for people's daily travel. According to data from the Ministry of Industry and Information Technology, by the end of 2018, the number of non-motor vehicles in the country reached 200 million. Since the development of urban traffic cannot meet the rapidly growing number of non-motorized vehicles, this has brought enormous pressure to urban traffic: such as non-motorized vehicles running red lights, non-motorized vehicles driving in violation of regulations, etc. In recent years, non-motor vehicle accidents have occurred frequently, with an annual growth rate ...

AI Technical Summary

Problems solved by technology

[0004] At present, the non-motor vehicle license plate image positioning technology still faces many difficulties, such as: the ambient light changes drastically, the daylight is strong, the night is weak, the front light and the backlight are different, the license plate image captured may contain more noise, or The image contrast is not strong, and the image is blurred and distorted due to motion, etc.

Just because it is a difficult problem to locate non-motor vehicle license plates in complex scenes, it cannot be effectively solved by only one processing method

Images