Systems and methods for automated, rapid detection of high-atomic-number materials

• US 8,633,823 B2
• Filed: 05/16/2010
• Issued: 01/21/2014
• Est. Priority Date: 05/16/2009
• Status: Active Grant

• Current Assignee: Rapiscan Systems Incorporated (OSI Systems Incorporated)
• Original Assignee: Rapiscan Systems Incorporated (OSI Systems Incorporated)
• Inventors: Chang, William, Franco, Edward D., Bendahan, Joseph, Browne, Jolyon A., Armistead, Robert A. Jr.
• Primary Examiner(s): Wu, Daniel
• Assistant Examiner(s): Barakat, Mohamed
• Application Number:
  US12/780,910
• Publication Number:
  US 20100295689A1
• Time in Patent Office: 1,346 Days
• Field of Search: 340/600, 340/539.26, 378/57, 348/80, 250/307, 250/310, 250/311, 250/308, 382/278, 382144-151, 382/171, 382/172
• US Class Current: 340/600
• CPC Class Codes: G01N 23/04 and forming images of the material
• CPC Class Codes: G01V 5/22 Active interrogation, i.e. by irradiating objects or goods using external radiation sources, e.g. using gamma rays or cosmic rays
• CPC Class Codes: G01V 5/222 measuring scattered radiation
• CPC Class Codes: G01V 5/224 Multiple energy techniques using one type of radiation, e.g. X-rays of different energies
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06V 2201/05 Recognition of patterns representing particular kinds of hidden objects, e.g. weapons, explosives, drugs

Abstract

The present invention is directed to an inspection system that has a radiation source, a detector array, an inspection region, and a processing unit, where the processing unit a) obtains a radiographic image, b) segments the radiographic image based on radiation attenuation or transmission, c) identifies at least one segmented area on the radiographic image, d) filters the at least one segmented area using at least one geometric filter, e) generates feature vectors using the filtered segmented area; and f) compares the feature vectors against predefined values to determine whether a high-atomic-number object is present.

Claims

1. An inspection system comprising:

a. a radiation source;

b. a detector array;

c. an inspection region bounded by said radiation source and detector array;

d. a processing unit, wherein, through operation of at least one processor, at least one memory, and programmatic instructions, said processing unit

i. obtains data representative of a radiographic image;

ii. segments said data based on radiation attenuation or transmission;

iii. identifies at least one segmented area within said data representative of said radiographic image;

iv. after identifying said at least one segmented area, filters said at least one segmented area using at least one geometric filter;

v. after filtering said at least one segmented area, generates a plurality of feature vectors using said filtered segmented area; and

vi. compares said feature vectors against predefined values to determine whether a high-atomic-number object is present.

2. The inspection system of claim 1 wherein said radiographic image has a spatial resolution of at least 25% of a minimum size of a threat object.

3. The inspection system of claim 1 wherein said radiation source is at least one of X-ray or gamma-ray radiation.

4. The inspection system of claim 1 wherein said processing unit generates a map of said segmented data representative of said radiographic image by determining local maximum peak attenuation values.

5. The inspection system of claim 1 wherein said processing unit generates a map of said segmented data representative of said radiographic image by determining minimum transmission values and applying edge gradient calculations.

6. The inspection system of claim 1 wherein said geometric filter is at least one of shape, symmetry, size or homogeneity.

7. The inspection system of claim 6 wherein said size filter is applied to at least one segmented area to identify dimensions selected on the basis of spatial resolution or penetration.

8. The inspection system of claim 6 wherein said shape filter is applied to at least one segmented area to identify a spatial aspect ratio of less than 20.

9. The inspection system of claim 6 wherein said segmented area has a first number of pixels and wherein a second defined area has a second number of pixels and wherein the homogeneity filter is applied to determine a ratio of said first number of pixels to said second number of pixels.

10. The inspection system of claim 1 wherein said processing unit generates a plurality of feature vectors using said filtered segmented area by:

a. Obtaining an image of said filtered segmented area;

b. Estimating background attenuation around said filtered segmented area;

c. Subtracting the background attenuation from said filtered segmented area to generate a net attenuation of the filtered segmented area;

d. Estimating dimensions of the area of interest using said net attenuation of the filtered segmented area;

e. Calculating an attenuation of the filtered segmented area as if it were a high-Z material; and

f. Comparing the calculated attenuation of the filtered segmented area of interest to the net attenuation of the filtered segmented area.

11. The inspection system of claim 1 wherein said feature vectors comprise at least one of maximum attenuation, net attenuation, a ratio of attenuation to an area of a suspicious object, a gradient of a suspicious object along a boundary, and a difference produced between measured background corrected attenuation and calculated attenuation.

12. The inspection system of claim 1 wherein said feature vectors are compared against predefined values to determine whether a high-atomic-number object is present.

13. An inspection system comprising a processing unit, wherein said processing unit comprises a plurality of programmatic instructions and wherein, when executed, said programmatic instructions:

i. segment data of a first radiographic image, which is representative of a first view of an object, and a second radiographic image, which is representative of a second view of the object, based on radiation attenuation or transmission;

ii. after segmenting said first and second radiographic images, filter at least one segmented area using at least one geometric filter for each of said images;

iii. after filtering said at least one segmented area, generate a plurality of feature vectors using said filtered segmented area for each of said images; and

iv. determine whether a high-atomic-number object is present using said feature vectors for each of said images.

14. The inspection system of claim 13 wherein each of said radiographic images is produced using at least one of an X-ray or gamma-ray radiation source.

15. The inspection system of claim 13 wherein said processing unit activates an alarm if each of said images indicates the presence of a high atomic number object.

16. The inspection system of claim 13 wherein said filter is at least one of shape, symmetry, size or homogeneity.

17. The inspection system of claim 16 wherein said size filter filters said at least one segmented area to identify dimensions selected on the basis of the inspection system'"'"'s spatial resolution or penetration.

18. The inspection system of claim 16 wherein said shape filter filters said at least one segmented area to identify a spatial aspect ratio of less than 20.

19. An inspection system comprising a processing unit, wherein said processing unit comprises a plurality of programmatic instructions and wherein, when executed, said programmatic instructions:

i. segment data of a first radiographic image, which is representative of an object and generated at a first energy level, and a second radiographic image, which is representative of the object and generated at a second energy level;

ii. after segmenting said first and second radiographic images, filter at least one segmented area using at least one geometric filter for each of said images;

iii. after filtering said at least one segmented area, generate a plurality of feature vectors using said filtered segmented area for each of said images;

iv. performs a ratio operation on said plurality of feature vectors, resulting in a ratio feature vector; and

v. determine whether a high-atomic-number object is present using said ratio feature vector.

20. The inspection system of claim 19 wherein said processing unit activates an alarm if each of said images indicates the presence of a high atomic number object.