Efficient license plate positioning method of convolutional neural network

Abstract

The invention provides an efficient convolutional neural network license plate positioning method, which mainly solves special problems such as license plate recognition, a network structure is optimized, an efficient convolutional neural network structure is provided, and the network structure has the advantages that the weight file size of the model can be reduced and the detection time can be shortened under the condition of not losing the detection accuracy. The method comprises the following steps: establishing a license plate database; generating an anchor box by using a K-means clustering method; designing an efficient deep neural network structure which is small in calculation amount and small in weight parameter number; training a network model on the final data set by using an Adam optimization algorithm; and a YOLOv3 is adopted as a comparison algorithm to evaluate the model. According to the convolutional neural network structure for license plate detection provided by theinvention, the weight file size of the model can be reduced and the license plate detection time can be reduced under the condition that the model detection accuracy is basically unchanged.

Description

technical field [0001] The invention belongs to the field of image recognition, and relates to an efficient license plate location method based on a deep convolutional neural network. Background technique [0002] With the development of social economy, automobiles have become an important way for people to travel daily. The types and quantities of automobiles are increasing rapidly, which puts forward higher requirements for traffic control. In recent years, intelligent processing technology has played an important role in many fields, and intelligent transportation systems have also emerged, which has greatly improved management efficiency and saved a lot of manpower. The license plate is an important identification of the vehicle, and each vehicle has a unique "identity certificate", which provides a strong guarantee for the unified management of the vehicle. Under the requirements of efficient vehicle management, automatic collection and recognition of license plates ha...

AI Technical Summary

Problems solved by technology

However, there are too many weight parameters in the network model, which causes the model weight file obtained after training to be too large, reaching 235M, which will lead to a decrease in the detection speed of the algorithm model, resulting in excessively high requirements for detection hardware equipment. The hardware requirements will cause the model to be too restrictive when deployed, which will lead to limited applicable scenarios

Images