Method and device for vehicle detection

Abstract

The present invention relates to a method and device for vehicle detection. The method comprises: after receiving a picture to be detected containing vehicle information, extracting the basic feature information of the picture to be detected by a predetermined algorithm; inputting the basic feature information into the pre-training generated In the And-Or model of the model, each level node is obtained, and each level node is output as a key node; the key nodes are associated, and the associated key nodes of each level are used as a better calculation branch; each calculation branch The hierarchical key nodes are converted into position parameters in the picture to be detected, and the graphic templates corresponding to the key nodes of each level in the calculation branch are determined according to the predetermined association relationship between the key nodes of each level and the graphic template; The position parameters and graphic templates corresponding to the nodes obtain and output the vehicle position information and vehicle layout relationship in the picture to be detected. The invention can efficiently identify vehicle information in complex scene pictures.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a vehicle detection method and device. Background technique [0002] At present, the identification of vehicle information is generally realized by detecting the target object in the vehicle information picture through an automated supervision system, such as detecting the license plate in the vehicle information picture. However, due to the diversity of real vehicle scenes, the irregular occlusion relationship between vehicles or the uncontrollable ratio of visible parts, the current recognition of vehicle information often encounters more interference and the recognition effect is not good. In addition, the traditional recognition of vehicle information is often carried out using a simple mode of manually setting features, and the efficiency of the recognition work is low when dealing with some complex scenes. Contents of the invention [0003] The technical problem...

AI Technical Summary

Problems solved by technology

However, due to the diversity of real vehicle scenes, the irregular occlusion relationship between vehicles or the uncontrollable ratio of visible parts, the current recognition of vehicle information often encounters more interference, and the recognition effect is not good.

In addition, the traditional identification of vehicle information is often carried out using a simple mode of manually setting features, and the efficiency of identification work is low when dealing with some complex scenes

Images