A license plate detection method, camera and server

Abstract

The embodiments of the present invention provide a license plate detection method, a camera and a server. The license plate detection method includes the following steps that: N target character segmentation blocks of a license plate image to be detected are obtained; a plurality of features are calculated for each of the target character segmentation blocks; on the basis of an algorithm according to the confidence of being contained in preset character segmentation blocks of abnormal license plate characters as for each kind of feature is calculated, confidence corresponding to each kind of features of each of the target character segmentation blocks is obtained through calculation, and a confidence vector is generated, and all generated confidence vectors are inputted into a pre-built second-layer license plate character prediction sub-model, and a prediction result which indicates whether each of the target character segmentation blocks contains the abnormal license plate characters can be obtained; and if the prediction result indicates that at least one of the target character segmentation blocks contains the abnormal license plate characters, it is determined that a license plate corresponding to the license plate image to be detected is an abnormal license plate, otherwise, the license plate corresponding to the license plate image to be detected is an normal license plate. With the license plate detection method of the invention adopted, the accuracy of license plate detection is improved.

Description

technical field [0001] The invention relates to the technical field of intelligent transportation, in particular to a license plate detection method, a camera and a server. Background technique [0002] With the continuous development of image processing, pattern recognition and artificial intelligence technology, intelligent transportation technology has shown great advantages in my country's transportation field, and the license plate detection system is the most critical link in the realization of intelligent transportation. Including camera and server two major hardware. In practical applications, the license plate detection system for abnormal license plate detection is particularly important. [0003] Generally, the detection of abnormal license plates may include two steps: segmentation of license plate characters and prediction of license plate characters. In the prior art, such a segmentation method of license plate characters is provided: extract the LBP texture f...

AI Technical Summary

Problems solved by technology

[0004] Obviously, although the above-mentioned detection method for abnormal license plate is feasible, in practical application, only by extracting a single image feature to segment the license plate characters, it is easy to have the problem of inaccurate character segmentation block position. The prediction based on single texture feature training When the model predicts abnormal characters, it is prone to inaccurate predictions, which leads to low detection accuracy of abnormal license plates

Images