All-class license plate recognition method based on object detection

Abstract

The invention belongs to the field of image processing and depth learning, and in particular provides a method for recognizing all kinds of vehicle license plates based on object detection, which is mainly used for recognizing all kinds of vehicle license plates under complex scenes. The method of the invention takes an object detection algorithm in the depth learning technology as the cut-in point, unifies the sub-tasks in the license plate recognition technology under the frame of the target detection, and firstly uses the character detection based on the depth learning and combines the character sorting to complete the character sequence recognition of the whole kind of license plate. Compared with the existing license plate recognition methods based on the traditional method and the depth learning method, the method of the invention has extremely strong universality, can be applied to all kinds of license plate recognition under complex scenes, and has very high recognition rate.

Description

technical field [0001] The invention belongs to the field of image processing and deep learning, and specifically provides a method for recognizing all types of license plates based on target detection, which is mainly used for recognizing all types of license plates in complex scenes. Background technique [0002] License plate recognition technology is a very important technology in intelligent transportation systems. Traditional license plate recognition technology generally includes three parts: license plate location, character segmentation, and character recognition; although license plate recognition algorithms have been widely used in parking lots, highways, However, there are still many problems waiting to be solved, such as the low overall recognition rate in complex scenarios such as highway monitoring, and the algorithm needs to be modified for new types of license plates. [0003] With the introduction of Alexnet in 2012, deep learning technology shines in image...

AI Technical Summary

Problems solved by technology

For the field of license plate recognition, the introduction of deep learning is a big trend, but most of the methods currently proposed have common shortcomings: they do not completely abandon the image processing steps that cause poor robustness in traditional methods; for example, the application numbers are: CN201710188171, The title of the invention is a kind of morphological processing and contour extraction in the invention patent of a license plate location method and device based on a full convolutional network. Another example is the second part of the invention patent with the application number CN201710432232 and the title of the invention is an efficient and accurate license plate recognition method. Value and projection operations; in addition, these methods only consider the recognition of several or even only one type of license plate. Once a new type of license plate is to be recognized, it is no longer applicable

Images