Method and system of inspecting vehicle

Patent metadata

Field	Value
Publication number	US-12417524-B2
Application number	US-202117928166-A
Country	US
Kind code	B2
Filing date	May 24, 2021
Priority date	May 28, 2020
Publication date	Sep 16, 2025
Grant date	Sep 16, 2025

Abstract

A method of inspecting a vehicle includes: acquiring a to-be-inspected image of an inspected vehicle (S 11 ); acquiring a visual feature of the to-be-inspected image using a first neural network model (S 12 ); retrieving a template image from a vehicle template library based on the visual feature of the to-be-inspected image (S 13 ); determining a variation region between the to-be-inspected image and the template image (S 14 ); and presenting the variation region to a user (S 15 ). The system of inspecting a vehicle includes a radiation imaging device ( 150 ), a display device ( 130 ), an image processor ( 140 ), and a storage device ( 120 ). The present disclosure further includes a computer-readable storage medium.

First claim

What is claimed is:

1. A method of inspecting a vehicle, comprising:
performing X-ray scanning on an inspected vehicle to obtain an X-ray image of the inspected vehicle;
preprocessing the X-ray image and performing a linear transformation on the preprocessed X-ray image to obtain a to-be-inspected image of the inspected vehicle;
acquiring a visual feature of the to-be-inspected image using a first neural network model;
retrieving a template image from a vehicle template library based on the visual feature of the to-be-inspected image;
determining a variation region between the to-be-inspected image and the template image;
marking a contour of the variation region with a color line on the to-be-inspected image according to data of the variation region, so as to prompt a position of a detected suspect in the inspected vehicle; and
presenting the to-be-inspected image with the position of the detected suspect in the inspected vehicle being prompted,
wherein the determining a variation region between the to-be-inspected image and the template image comprises: registering the to-be-inspected image with the template image; extracting feature maps of the registered to-be-inspected image and the registered template image; and determining a difference between the feature maps of the to-be-inspected image and the template image,
wherein the feature maps of the to-be-inspected image and the template image are extracted using a second neural network model.

2. The method according to claim 1 , wherein the vehicle template library is constructed by:
acquiring a plurality of template vehicle images;
acquiring visual features of the plurality of template vehicle images using the first neural network model; and
clustering the visual features, and determining the clustered visual features as the vehicle template library.

3. The method according to claim 2 , wherein the retrieving a template image from a vehicle template library based on the visual feature of the to-be-inspected image comprises:
calculating feature similarities between the visual feature of the to-be-inspected image and all visual features in the vehicle template library, and determining a template vehicle image corresponding to a maximum feature similarity as the template image.

4. The method according to claim 1 , wherein the presenting the variation region to a user comprises:
highlighting the variation region on the to-be-inspected image.

5. A non-transitory computer-readable storage medium storing instructions, wherein the instructions, when executed by a processor, cause the processor to execute the method according to claim 1 .

6. A system of inspecting a vehicle, comprising:
a display device,
an image processor, and
a storage device coupled with the image processor and storing computer-readable instructions, wherein the computer-readable instructions, when executed by the image processor, cause the image processor to:
perform X-ray scanning on an inspected vehicle to obtain an X-ray image of the inspected vehicle;
preprocess the X-ray image and perform a linear transformation on the preprocessed X-ray image to obtain a to-be-inspected image of the inspected vehicle;
acquire a visual feature of the to-be-inspected image using a first neural network model;
retrieve a template image from a vehicle template library based on the visual feature of the to-be-inspected image;
determine a variation region between the to-be-inspected image and the template image;
mark a contour of the variation region with a color line on the to-be-inspected image according to data of the variation region, so as to prompt a position of a detected suspect in the inspected vehicle; and
present the to-be-inspected image with the position of the detected suspect in the inspected vehicle being prompted,
wherein the computer-readable instructions, when executed by the image processor, further cause the image processor to: register the to-be-inspected image with the template image; extract feature maps of the registered to-be-inspected image and the registered template image; and determine a difference between the feature maps of the to-be-inspected image and the template image,
wherein the feature maps of the to-be-inspected image and the template image are extracted using a second neural network model.

7. The system according to claim 6 , wherein the vehicle template library is constructed by:
acquiring a plurality of template vehicle images;
acquiring visual features of the plurality of template vehicle images using the first neural network model; and
clustering the visual features, and determining the clustered visual features as the vehicle template library.

8. The system according to claim 7 , wherein the computer-readable instructions, when executed by the image processor, further cause the image processor to:
calculate feature similarities between the visual feature of the to-be-inspected image and all visual features in the vehicle template library, and determine a template vehicle image corresponding to a maximum feature similarity as the template image.

9. The system according to claim 6 , wherein the computer-readable instructions, when executed by the image processor, further cause the image processor to:
control the display device to highlight the variation region on the to-be-inspected image.

Assignees

• Nuctech Co Ltd
• Univ Tsinghua

Inventors

• Chen Zhiqiang
• Zhang Li
• Li Yuanjing
• Xing Yuxiang
• Shi Jie

Classifications

• G06T2207/30248

  Vehicle exterior or interior · CPC title

• G06T2207/20084

  Artificial neural networks [ANN] · CPC title

• G06T2207/10116

  X-ray image · CPC title

• G06T1/20

  Processor architectures; Processor configuration, e.g. pipelining · CPC title

• G06T1/0007

  Image acquisition · CPC title