A Coarse License Plate Location Method Based on Adaptive Edge Features

Abstract

The invention provides a rough license plate positioning method based on adaptive edge features, comprising the following steps: grayscale image blurring processing; based on density information of vertical edges, using adaptive edge features to dynamically obtain the distance threshold between adjacent points in a row , connect adjacent points smaller than the threshold to form a connected area; based on the basic feature information of the connected area, including height, width, aspect ratio, and area, remove the non-connected area of ​​the license plate; according to the edge information around the candidate connected area, Corresponding expansion is carried out to form a license plate candidate rectangular area. The present invention can obtain more accurate connected areas and exclude more non-license plate areas. At the same time, for images containing license plates of different sizes, the entire processing process does not require scale transformation, and the corresponding positioning detection can be completed, and the algorithm is more efficient.

Description

technical field [0001] The invention relates to the technical field of license plate recognition, in particular to a rough license plate positioning method based on adaptive edge features. Background technique [0002] License plate recognition is the core technology of intelligent transportation, which includes three parts: license plate location, character segmentation, and character recognition. Among them, the license plate positioning is the most important part of the whole technology. The quality of the license plate positioning directly affects the subsequent steps, which in turn affects the overall recognition speed and accuracy. [0003] License plate location refers to quickly find out the approximate area where the license plate is located in an image. It is generally realized by combining coarse positioning and fine positioning. The specific method is: rough positioning quickly excludes most of the non-license plate areas in the entire image, leaving candidate ...

AI Technical Summary

Problems solved by technology

But there are also a series of disadvantages: since the distance threshold between the morphological structural elements and the adjacent points of the row is mostly fixed, the adaptability to license plates of different sizes is poor, which leads to more false detections.

Some algorithms adopt the method of multiple detection after scale transformation. Although the effect of adapting to license plate images of different sizes is achieved, the repeated operation process greatly reduces the speed of the algorithm and seriously affects the final performance.

Images