Vehicle Brand Model Recognition Method Based on Rapid Learning Framework

Abstract

The invention discloses a vehicle brand model recognition method based on a rapid learning framework, comprising the following steps: building a spark platform, acquiring a monitoring video stream; detecting a license plate from the acquired video image, acquiring a vehicle face image according to the position of the license plate, and performing vehicle Face recognition; expand up, down, left, and right according to the width and height of the license plate to obtain a car face image; construct a multi-scale car face image, divide it into overlapping blocks, extract local binary and gradient direction histogram features for each small block, and Combine into the final car face features; train multi-level cascaded classifiers, the first level adopts multi-classifier voting, when most classifiers vote for the same target, accept the result, otherwise enter the next level; second The integrated classification system is used in the first stage, and multiple sub-classifiers are trained through feature mapping, and the results of each classifier are fused to obtain the final classification result. It can accurately, quickly and stably identify vehicle brands and specific models in massive amounts of data.

Description

technical field [0001] The invention relates to the fields of image pattern recognition and intelligent transportation, in particular to a vehicle brand and model recognition method based on a rapid learning framework. Background technique [0002] With the continuous development of the national economy, motor vehicles have become an indispensable means of transportation in people's daily life. But the ensuing traffic problems are also becoming more and more prominent. All countries in the world have increased investment in the management of traffic systems, and gradually formed the research field of road traffic management. [0003] Vehicle type recognition is the main task and key technology in intelligent transportation systems, and it has a wide range of applications, such as automatic highway billing, parking lot management, inspection of stolen vehicles, management of licensed vehicles, etc. The difficulty of vehicle identification lies in the universal adaptability ...

AI Technical Summary

Problems solved by technology

Although this method uses self-learning features, it uses a classifier, and the classification effect is poor.

Moreover, when faced with massive data, the processing speed is slow and the accuracy rate is relatively low.

Images