License plate self-correcting identification method based on license plate credibility

Abstract

The invention belongs to the field of credibility license plate identification, and specifically relates to a license plate self-correcting identification method based on license plate credibility. The method includes the steps of: judging the provided credibility of a license plate identification result, setting initial credibility threshold values and providing default province Chinese character settings; obtaining influence of the credibility on accuracy after adjustment, and finding out T1; obtaining an identification result whose credibility is higher than T1, and then finding out a Chinese character whose error is easy to identify for replacement; on the basis of T1, obtaining the initial credibility threshold values, and correcting a province Chinese character in a license plate number whose credibility is between T1 and T2 into the abovementioned default province Chinese character; and then finding out the credibility threshold value T2 of a point with the largest influence, and obtaining a final corrected result. The method can realize further optimization of the identification result on the basis of a result provided by a license plate identification algorithm at present, so as to further improve accuracy of the identification result, and identification reliability can also be effectively guaranteed.

Description

technical field [0001] The invention belongs to the field of license plate recognition with credibility, and in particular relates to a license plate self-correcting recognition method based on the credibility of the license plate which is applied on the basis of the current license plate recognition algorithm to further improve the accuracy of the recognition algorithm. Background technique [0002] Intelligent Transportation System (Intelligent Transportation System, referred to as ITS) is the development trend of road traffic management in the 21st century. The vehicle license plate is the unique identity of the vehicle. As an important part of the intelligent transportation system, the vehicle license plate recognition system (LPRS) has attracted the attention of experts in recent years, and the products have been put into use in order to improve the efficiency and level of traffic management and save Scientific and standardized management of manpower, material resources...

AI Technical Summary

Problems solved by technology

When necessary, it is even necessary to manually identify the license plate number with too low reliability in the later stage to improve the identification accuracy of the overall data, which is labor-intensive; of course, the "low" reliability here The water dividing line is also determined by experience

Especially in my country, there are a large number of low-resolution cameras with a camera resolution of only 720P in traffic management, and the capture of license plates is relatively blurred, especially the Chinese characters; and some license plates have a high inclination angle, which further leads to extremely low resolution of the photos

As far as the current vehicle license plate recognition results are concerned, it is naturally classified as a low-confidence category, which further reduces the recognition accuracy of the vehicle license plate recognition system.

Images