Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-task-convolutional-neural-network-based detection method of fake-licensed car

A technology of convolutional neural network and detection method, which is applied in the direction of neural learning method, biological neural network model, neural architecture, etc., and can solve the problems of narrowing the scope of investigation, slow efficiency, and low recognition accuracy

Inactive Publication Date: 2018-08-28
ZHEJIANG UNIV OF TECH
View PDF9 Cites 26 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0013] In order to overcome the shortcomings of low recognition accuracy, slow efficiency and inability to identify high imitation license plate vehicles in the existing detection and identification methods for license plates, the present invention provides a detection method with high accuracy, fast efficiency, and capable of coping with high imitation license plates , relying on the layer-by-layer comparison technology, from the macroscopic characteristics of the vehicle to the microscopic characteristics of the vehicle, gradually improve the accuracy of detection and reduce the scope of investigation

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-task-convolutional-neural-network-based detection method of fake-licensed car
  • Multi-task-convolutional-neural-network-based detection method of fake-licensed car
  • Multi-task-convolutional-neural-network-based detection method of fake-licensed car

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0072] The present invention will be further described below in conjunction with the accompanying drawings.

[0073] refer to Figure 1 ~ Figure 4 , a method for detecting license plates based on a multi-task convolutional neural network, comprising the following steps:

[0074] 1) Using time-space constraints to process the acquired vehicle image data to obtain suspected fake vehicles. The suspected fake vehicles refer to suspicious vehicles detected by the screening system. The car series and body color are the same or highly similar;

[0075] 2) Use Faster R-CNN to locate and segment the image of the car body;

[0076] 3) Realize convolution sharing for the first 5 layers of convolution;

[0077] 4) Training the positioning network of the car face feature and the network for extracting the private face feature of the public face feature;

[0078] 5) Feature extraction is performed on the public face part and the private face part of the suspected license plate vehicle r...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

Provided in the invention is a multi-task-convolutional-neural-network-based detection method of a fake-licensed car. A suspected fake-licensed car is obtained by using time-space constraints; a car front-face part image is located and segmented by using Faster R-CNN; features of car public face parts being basic features of the car of the suspected fake-licensed car are compared; and then subtlefeatures of a car private face part being a car inspection identifier of the high-imitation fake-licensed car are detected and compared. Therefore, problems that low accuracy and difficult problem carlocking are caused by only utilization of the time-space constraint analysis as well as the high-imitation fake-licensed car can not be identified purely based on the car public face are solved; andthe accuracy and efficiency are high. The multi-task-convolutional-neural-network-based detection method is suitable for high-imitation fake-licensed cars.

Description

technical field [0001] The invention relates to an analysis method, in particular to the application of technologies such as computer vision, digital image processing, pattern recognition, deep learning and deep convolutional neural network in the field of detection and recognition of licensed vehicles. Background technique [0002] With the development of society and economy, in recent years, the number of cars has increased sharply, and a series of traffic problems have become increasingly prominent. Refers to vehicles that go on the road through forged or illegally obtained other vehicle license plates and driving permits. Because the owner of the licensed car rarely uses traffic laws to restrain himself when driving on the road, the resulting traffic accidents, legal disputes, and smuggling crimes will definitely bring great instability to the society. Although public security, traffic control and other departments spent a lot of manpower and financial resources on rect...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06K9/00G06K9/46G06N3/04G06N3/08
CPCG06N3/08G06V20/584G06V10/462G06N3/045
Inventor 陈朋汤一平王丽冉何霞袁公萍
Owner ZHEJIANG UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products