Systems and methods for inspecting cargoes

Patent metadata

Field	Value
Publication number	US-10074166-B2
Application number	US-201514700249-A
Country	US
Kind code	B2
Filing date	Apr 30, 2015
Priority date	Dec 30, 2014
Publication date	Sep 11, 2018
Grant date	Sep 11, 2018

Abstract

The present disclosure discloses a method and system for inspecting cargoes. The method comprises: acquiring a transmission image of the inspected cargoes; processing the transmission image to acquire an interested region; extracting features from the interested region, and determining cargo information of the inspected cargoes according to the extracted features; and providing a proposed treatment suggestion of the cargoes based on the determined cargo information and at least a part of information in a manifest. The above solution can facilitate an image judgment person to accurately judge whether the concerned cargoes are allowed to pass.

First claim

We claim:

1. A method for inspecting cargoes, comprising:
subjecting inspected cargoes to a radiation;
acquiring a transmission image of the inspected cargoes from transmitted rays of the radiation;
processing the transmission image to acquire an interested region;
extracting features from the interested region, and determining cargo information of the inspected cargoes according to the extracted features; and
providing a proposed treatment suggestion of the cargoes based on the determined cargo information and at least a part of information in a manifest,
wherein the cargo information is spatial arrangement of the cargoes, and
the step of extracting features from the interested region comprises extracting description information of the spatial arrangement of the inspected cargoes from the interested region; and
the step of providing a proposed treatment suggestion of the cargoes based on the determined cargo information and at least a part of information in a manifest comprises: determining a consistency probability by multiplying the probability of a type of cargoes determined based on real spatial arrangement from the manifest and the conditional probability of the real spatial arrangement based on a predicted one from the transmission image, and in a case that the consistency probability is lower than a predetermined value, judging that the spatial arrangement of the cargoes are inconsistent with the type of the cargoes in the manifest.

2. The method according to claim 1 , wherein the subjecting inspected cargoes to a radiation comprises subjecting the inspected cargoes to an x-ray radiation.

3. The method according to claim 1 , wherein the step of extracting description information of the spatial arrangement of the inspected cargoes from the interested region comprises:
partitioning the interested region and extracting edges from the image of the region;
extracting textual features in effective positions of the extracted edges; and
finding elements nearest to the textual features from a dictionary, to obtain histogram description features of the region, wherein the dictionary is established based on training image samples.

4. The method according to claim 3 , wherein the dictionary is established by
sampling each sampling image at regular intervals, and extracting a texture feature in each position; and
forming the dictionary by aggregating the textual features extracted in all positions, to acquire feature description of multiple aggregation centers.

5. The method according to claim 1 , wherein the interested region is determined by extracting a cargo region of the inspected cargoes and removing regions without cargoes according to a threshold.

6. The method according to claim 1 , further comprising:
prompting to a user a cargo region which deviates from a main body spatial arrangement, a small cargo region for which a type of cargoes cannot be predicted, or a local region of which a texture is significantly different from the global texture.

7. The method according to claim 1 , further comprising:
determining a type of cargoes which are placed individually in a small pile and cannot involve in global information deduction using an interactive deduction manner.

8. The method according to claim 1 , further comprising:
deducing a cargo type distribution which can present such textures, an average atomic number, and a position in the container according to texture information of a local region.

9. The method according to claim 8 , further comprising:
extracting texture features in a specified region;
acquiring a list of texture blocks with similarities larger than a predetermined value based on image content retrieval; and
establishing the cargo type distribution which presents such textures by determining cargo type information of the list of texture blocks.

10. The method according to claim 9 , wherein the average atomic number of the inspected cargoes is acquired by using a dual-energy material recognition algorithm.

11. The method according to claim 9 , wherein an actual position of the specified region in the container is acquired according to the scaling of the image and a pixel position of the specified region.

12. A system for inspecting cargoes, comprising:
a radiation imaging system configured to scan the inspected cargoes to acquire a transmission image of the inspected cargoes; and
a data processing apparatus configured to process the transmission image to acquire an interested image, extract features from the interested region, determine cargo information of the inspected cargoes according to the extracted features, and provide a proposed treatment suggestion of the cargoes based on the determined cargo information and at least a part of information in a manifest, wherein the cargo information is spatial arrangement of the cargoes, and
the data processing apparatus is further configured to extract features from the interested region by extracting description information of the spatial arrangement of the inspected cargoes from the interested region; and
the step of providing a proposed treatment suggestion of the cargoes based on the determined cargo information and at least a part of information in a manifest comprises—: determining a consistency probability by multiplying the probability of a type of cargoes determined based on real spatial arrangement from the manifest and the conditional probability of the real spatial arrangement based on a predicted one from the transmission image, and in a case that the consistency probability is lower than a predetermined value, judging that the spatial arrangement of the cargoes are inconsistent with the type of the cargoes in the manifest.

Assignees

• Univ Tsinghua
• Nuctech Co Ltd

Inventors

• Kang Kejun
• Chen Zhiqiang
• Zhang Li
• Zhao Ziran
• Liu Yaohong
• Zhang Jian
• Gu Jianping
• Li Qiang
• Zhang Duokun

Classifications

• G06T7/0004 Primary

  Industrial image inspection · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06K9/6215

  Physics · mapped topic

• G06K2009/4666

  Physics · mapped topic