Terahertz image recognition method and system based on deep learning

A technology of image recognition and deep learning, applied in the field of terahertz image recognition, to achieve efficient detection process, reduce the time of manual image judgment, and reduce the number of effects

Inactive Publication Date: 2019-10-29
博微太赫兹信息科技有限公司
View PDF8 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is: how to automatically detect and recognize the dangerous target carried by the human body in the terahertz image, and provides a terahertz image recognition method based on deep learning, which realizes an automatic and efficient detection process, Effectively improves the speed and accuracy of security inspection, can greatly reduce the time for manual map judgment, and can reduce the number of staff; it can effectively filter a large amount of information, retain valid information, eliminate redundant alarm information, and complete and concise detection results display, effectively reducing the workload of staff

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
  • Terahertz image recognition method and system based on deep learning
  • Terahertz image recognition method and system based on deep learning
  • Terahertz image recognition method and system based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0063] like figure 1 As shown, this embodiment provides a technical solution: a deep learning-based terahertz image recognition method, including the following steps:

[0064] S1: Design and train a convolutional neural network

[0065] Design a convolutional neural network, the backbone network structure of the convolutional neural network is the VGG16 network structure, and replace the fully connected layer of the VGG16 network structure with a convolutional layer, and add four convolutional layers to form the convolutional neural network. Network structure, and then train the convolutional neural network to obtain the weight file of the convolutional neural network;

[0066] S2: Detect dangerous target information in terahertz raw images

[0067] Receive the original terahertz image, use the weight file formed by the training of the convolutional neural network to detect the dangerous target in the original terahertz image, and obtain the information of the dangerous targ...

Embodiment 2

[0107] The terahertz image recognition method based on deep learning in this embodiment mainly includes two parts: a convolutional neural network and a structured detection target sequence.

[0108] like figure 2 As shown, it is a block diagram of the overall process of target detection. The overall process of target detection is as follows: collect a large number of terahertz original images formed by the detection equipment, first undergo image preprocessing, send them to the convolutional neural network for training, and obtain the weight file of the network; Load the weight file of the network in the detection device to detect dangerous targets on the original terahertz image; when the detection device is in use, receive the original terahertz image transmitted by the device, and the weight file formed by the convolutional neural network The dangerous target in the Hertz original image is detected; the detected target information, through the target information structurin...

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

The invention discloses a terahertz image recognition method and system based on deep learning, and belongs to the technical field of terahertz image recognition, and the method comprises the following steps: S1, designing and training a convolutional neural network; s2, detecting dangerous target information in the terahertz original image; s3, performing structured processing on the target information; and S4, displaying the image information. In the step S1, a weight file of the convolutional neural network is loaded into detection equipment before a dangerous target existing in the terahertz original image is detected. According to the method and system, an automatic and efficient detection process is realized, the security check speed and precision are effectively improved, the manualimage judgment time can be greatly reduced, and the number of workers can be reduced; a large amount of information can be effectively filtered, effective information is reserved, redundant alarm information is eliminated, a complete and concise detection result is displayed, and the workload of workers is effectively reduced.

Description

technical field [0001] The present invention relates to the technical field of terahertz image recognition, in particular to a method and system for terahertz image recognition based on deep learning. Background technique [0002] Terahertz wave is a kind of electromagnetic wave, and the human body also emits electromagnetic waves in the terahertz band. Through certain technical means, the terahertz wave emitted by the human body is detected, and the detected image is processed to form a terahertz image. [0003] At present, due to its advantages of no radiation, non-contact, and non-stop, terahertz human body security inspection is more and more popular among customers and accepted by the general public in the security inspection industry. Due to the special imaging principle of terahertz images, which are very different from optical images, users without certain training cannot obtain effective information. [0004] Due to the special imaging principle of terahertz equip...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/32G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V10/25G06V2201/07G06N3/045G06F18/24
Inventor 罗美林郭林胡其枫卜伟华张华坤唐红强查文锦汪潮
Owner 博微太赫兹信息科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap