Multiview x-ray based system for detecting contraband such as in baggage

• US 6,088,423 A
• Filed: 06/05/1998
• Issued: 07/11/2000
• Est. Priority Date: 06/05/1998
• Status: Expired due to Term

• Current Assignee: L-3 Communications Security and Detection Systems Corporation Delaware
• Original Assignee: Vivid Technologies, Inc. (Revvity, Inc.)
• Inventors: Krug, Kristoph D., Tortora, John O., Bijjani, Richard, Eilbert, Richard F.
• Primary Examiner(s): Porta, David P.
• Assistant Examiner(s): Ho, Allen C
• Application Number:
  US09/092,295
• Time in Patent Office: 767 Days
• Field of Search: 378/57, 378/51, 378/69, 378/9, 378/4
• US Class Current: 378/57
• 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 5/224 Multiple energy techniques using one type of radiation, e.g. X-rays of different energies

Abstract

A multiview X-ray based inspection system detects contraband such as explosives, weapons, drugs and money in packages such as baggage. Each bag moves through at least three stationary, fan-shaped beams of x-rays that are in parallel planes spaced from each other in the direction of bag travel, to produce three projection views of the bag taken from different angles. These projection views are processed to extract two-dimensional outlines of likely individual objects in the bag, and the two-dimensional outlines are matched between different projection views to extract three-dimensional outlines of likely objects in the bag. The system uses these three-dimensional outlines to estimate effective atomic numbers and mass densities of material in these objects, and then to determine the presence of likely contraband in the baggage by testing these estimates against windows in a multi-dimensional space in which one dimension is effective atomic number and another is mass density. The typical windows correspond to known contraband materials. The system can derive measurements from additional modalities such as nuclear quadrupole resonance (NQR) for use in conjunction with the x-ray measurements in finding likely contraband.

Claims

1. A multi view x-ray baggage inspection method for detecting contraband in baggage comprising:

moving baggage through at least three differently oriented x-ray beams which are spaced from each other in the direction of baggage movement through a tunnel having sides, said beams emanating from respective x-ray sources, each pair of sources being spaced from each other by at least a major portion of a side of the tunnel;

deriving dual energy digital x-ray data related to interactions between the baggage and the three x-ray beams;

processing the x-ray data with a computer to identify likely individual objects in the baggage and to produce digital estimates related to effective atomic number and digital estimates related to mass density of the likely objects; and

finding the likely presence of contraband in the baggage by testing the estimates relative to multi-dimensional space in which one dimension relates to density and another to effective atomic number.

2. A method as in claim 1 in which said moving comprises moving the baggage through three stationary x-ray beams.

3. A method as in claim 2 in which said moving comprises moving the baggage solely in a translational motion through the three stationary x-ray beams.

4. A method as in claim 3 in which said deriving comprises deriving dual energy x-ray data for each of said three x-ray beams.

5. A method as in claim 4 in which said deriving comprises deriving x-ray data forming three respective projection views of the baggage, one for each of the three stationary x-ray beams.

6. A method as in claim 5 in which said processing comprises identifying, in each of the three projection views, feature points or regions which are likely to conform to edges of the likely objects.

7. A method as in claim 6 in which said processing comprises finding properties of said feature points or regions.

21. A method as in claim 1 including subjecting the baggage to NQR inspection and using results of the NQR inspection to facilitate detection of contraband in the baggage.

22. A method as in claim 21 in which the processing further comprises processing of data derived from the NQR inspection and the finding of the likely presence of contraband comprises including estimates derived from said processing of data derived from the NQR inspection in said multi-dimensional space.

8. A multi view x-ray baggage inspection method for detecting contraband in baggage comprising:

moving baggage through at least three differently oriented x-ray beams which are spaced from each other in the direction of baggage movement;

deriving dual energy digital x-ray data related to interactions between the baggage and the three x-ray beams;

processing the x-ray data with a computer to identify likely individual objects in the baggage and to produce digital estimates related to effective atomic number and digital estimates related to mass density of the likely objects; and

finding the likely presence of contraband in the baggage by testing the estimates relative to multi-dimensional space in which one dimension relates to density and another to effective atomic number;

in which said processing comprises finding any tails of said feature points in a first direction which is transverse to the direction of baggage movement as well as in a second direction which is transverse to the first direction.

9. A method as in claim 8 in which said processing comprises chaining feature points or aggregating regions having compatible properties into likely two-dimensional outlines of likely objects in each of said projection views.

10. A method as in claim 9 in which said processing comprises identifying tuples having three-dimensional properties by matching said two dimensional outlines of likely objects in different projection views.

11. A method as in claim 10 in which said identifying of tuples comprises selecting tuples based on pairs of feature points which are from respective pairs of the two-dimensional outlines and meet selected criteria for matching.

12. A method as in claim 11 in which said processing comprises chaining tuples into likely three-dimensional outlines of likely objects.

13. A method as in claim 12 in which said chaining of tuples comprises forming a tuple chain for each likely object from each of a number of different slice views.

14. A method as in claim 13 in which said step of chaining tuples comprises using attributes of a tuple to form a polygon oriented transversely to the direction of baggage movement, said polygon matching a current estimate of the likely outer boundary of a likely object, and repeating for other tuples which are likely to belong to the same object.

15. A method as in claim 14 in which said processing comprises adjusting the shape of the polygon based on information from said projection views.

16. A method as in claim 15 in which said adjusting comprises adjusting the positions of a subset of vertices of a polygon from the same likely object in accordance with measurements for projection views, and adjusting the positions of other vertices for the same likely object based on adjusted positions of said subset of vertices.

17. A method as in claim 16 in which said finding the likely presence of contraband in the baggage comprises estimating average effective atomic numbers and average density for the likely objects in three-dimensional space in the baggage, and matching the estimates for the respective likely objects with windows in multi-dimensional space defined by at least effective atomic number and by density.

18. A method as in claim 17 including estimating dimensions of the baggage, and wherein the chaining of feature points and of tuples or aggregates of regions comprises taking into account said estimated dimensions.

19. A method as in claim 18 in which the producing of digital estimates related to mass density of a likely object comprises basing the estimates on measurements in the three projection views for positions of the vertices of polygons forming the likely three-dimensional outline of the object.

20. A method as in claim 19 in which additional image properties area used for the classifying of objects as contraband.

23. An x-ray baggage inspection method comprising:

moving baggage through at least three differently oriented, stationary fan-shaped x-ray beams which are parallel to each other;

deriving dual energy digital x-ray data related to interactions between the baggage and the x-ray beams;

computer-processing the x-ray data to identify likely three-dimensional individual objects in the baggage and to produce digital estimates related at least to effective atomic number and to mass density of the likely objects; and

finding the presence of likely contraband in the baggage by testing the estimates related at least to effective atomic number and mass to density of the likely individual objects in the baggage relative to multi-dimensional space in which density is one dimension and effective atomic number is another dimension.

24. A method as in claim 23 in which said moving comprises moving the baggage through only three x-ray beams.

25. A method as in claim 24 in which said moving comprises moving the baggage through the x-ray beams solely in translational motion relative to the beams.

26. A method as in claim 23 including deriving measurements from the baggage by a modality other than x-rays and using said measurements to facilitate the finding of likely contraband.

27. A method as in claim 26 in which said measurements are non-optical measurements.