Nuclear Material Detection System

Abstract

A method for automatically detecting nuclear material using radiographic images of a cargo container includes receiving a plurality of radiographic images of the cargo container and aligning the plurality of images with respect to each other to produce registered images. The method also includes segmenting the registered images using the atomic number and other edge/texture information in order to locate one or more regions of interest within the registered images and estimating atomic number information for each of a predetermined number of portions of the registered images. The method includes assigning a threat level and a confidence value to regions of interest identified as a potential threat and evaluating the regions of interest identified as potential threats using material context information and adjusting, based on the evaluation, the threat level values and confidences of the regions of interest identified as potential threats. The method also includes providing the regions of interest and adjusted threat level and confidence values as output to an operator station.

Description

[0001] The present application claims the benefit of U.S. Provisional Patent Application No. 60/940,632, entitled "Threat Detection System", filed May 29, 2007, which is incorporated herein by reference in its entirety.

[0002] Embodiments of the present invention relate generally to detection of nuclear (or other high atomic number) materials and, more particularly, to systems and methods for computerized automatic detection of nuclear materials in cargo conveyances.

[0003] In order to evaluate (or screen) cargo conveyances in an efficient manner it is desirable to automatically analyze images obtained by scanning the cargo conveyances. The images can be analyzed to determine whether a suspicious (or potential threat) material, such as a special nuclear material or other material having a high atomic number, may be present in the cargo container. In performing an automatic analysis of radiographic images to determine whether a potential threat material is present in the image, a need for sensitivity is often balanced against a need for a low false alarm rate. These competing needs are often expressed as requirements that an automatic image analysis system have a certain probability of detection of a potential threat material and a certain confidence of a true positive indication.

[0004] Embodiments of the automatic nuclear material detection method and system of the present invention may provide a reduced false alarm rate while maintaining a desired rate of detecting threats by increasing both the rate of identifying potential threats and the rate of identifying typical false alarms.

[0005] One exemplary embodiment includes a method for automatically detecting nuclear material using radiographic images of a cargo container includes receiving a plurality of radiographic images of the cargo container and aligning the plurality of images with respect to each other to produce registered images. The method also includes segmenting the registered images using the atomic number and other pattern information in order to locate one or more regions of interest within the registered images and estimating atomic number information for each of a predetermined number of portions of the registered images. The method includes assigning a threat level and a confidence value to regions of interest identified as a potential threat and evaluating the regions of interest identified as potential threats using material context information and adjusting, based on the evaluation, the threat level values and confidences of the regions of interest identified as potential threats. The method also includes providing the regions of interest and adjusted threat level and confidence values as output to an operator station.

[0006] Another exemplary embodiment includes a system for automatically detecting nuclear material in a radiographic image of a cargo conveyance. The system includes a material domain imaging module, an object segment recognition module, a material context analysis module, and an advanced cognitive arbitration module.

[0007] The material domain imaging module is adapted to receive a plurality of radiographic images and to determine an atomic number for each pixel in a combined atomic number image, the atomic number based on an analysis of a pixel in each of the radiographic images.

[0008] The object segment recognition module is coupled to the material domain imaging module and is adapted to perform an image segmentation operation on the combined atomic number image and to identify and return one or more object regions to the material domain imaging module, whereby the material domain imaging module can perform an object level atomic number analysis using a region of the combined atomic number image defined by the one or more object regions.

[0009] The material context analysis module is coupled to the material domain imaging module and is adapted to receive the combined atomic number image, having been segmented into the one or more object regions, and to generate a hypothesis as to whether an object region that has been assigned a high atomic number is in a suspicious location of the image, the hypothesis generation based on a computer-usable representation of expert knowledge and historical data, the material context analysis module is also adapted to provide the hypothesis as output to another module.

[0010] The advanced cognitive arbitration module is coupled to the material domain imaging module and is adapted to receive a plurality of threat hypotheses as inputs, each having an associated confidence value, the advanced cognitive arbitration module adapted to rank the inputs, assign a confidence value to each ranked input and provide a ranked list of threats, with associated confidence values, as output.

[0011] Another exemplary embodiment includes a threat detection system. The threat detection system includes means for determining estimated material atomic number values based on two or more radiographic images produced using a plurality of energy levels and generating an estimated atomic number based on the estimated material atomic number values. System is also capable of creating a hybrid image which contains gray level intensity information and atomic numbers by superimposing the atomic values onto the original gray level images measured using the radiographic system. The threat detection system also includes means for segmenting the hybrid image to define one or more regions of interest and means for analyzing a threat level of each region of interest using material context information. The threat detection system also includes means for arbitrating among multiple potential threat results to determine a final threat decision array for output.

BRIEF DESCRIPTION OF THE DRAWINGS

[0012] FIG. 1 is a block diagram of an object screening system including an exemplary nuclear material detection system;

[0013] FIG. 2 is a block diagram of an exemplary nuclear material detection system;

[0014] FIG. 3 is a block diagram of an exemplary nuclear material detection system showing interfaces between four processing modules; and

[0015] FIG. 4 is a flowchart of an exemplary method for detecting nuclear materials in a screened object.

DETAILED DESCRIPTION

[0016] In general, embodiments of the present invention use an image processing system to automatically identify and detect potential threats based on a set of radiographic images of an object being screened. The image processing system includes four primary functional areas: material Z value (or atomic number) determination; object segmentation; false alarm reduction; and advanced decision making. These four areas can be embodied as a material domain image (MDI) processor, object segmentation radiograph (OSR) processor, material context analysis (MCA) processor, and an advanced cognitive arbitrator (ACA) processor. These functional areas and processors can comprise computer hardware, software, or both.

[0017] The nuclear material detection system and method can be used to detect unauthorized, illegal or illicit attempts to import, assemble, or transport a nuclear explosive device (or a portion thereof), fissile material and/or radiological material. Such attempts may be connected with a potential threat to safety or security. These contraband materials or devices often can be detected and identified based on an estimate of the associated atomic number of the materials. For example, materials with an atomic number (or Z value) greater than 72 may be categorized as high atomic number materials. Elements with a high atomic number include special nuclear materials (SNM) such as plutonium (Pu), highly enriched uranium (HEU), and some elements (e.g., lead (Pb) and tungsten (W)) that may be effective in shielding SNM or other radioactive materials from passive gamma radiation detection.

[0018] In addition to automatically detecting materials having a high Z value, embodiments can include a capability for detecting traditional contraband such as drugs, currency, guns, and explosives. The capability to detect traditional contraband may be fully automatic or may include some manual operator image analysis.

[0019] FIG. 1 is a block diagram of an object screening system including an exemplary nuclear material detection system. In particular an object screening system 100 can be used to screen an object to be scanned 102 in order to detect contraband such as nuclear material. The object 102 is subjected to two or more electromagnetic energies (with two, 104a and 104b, being shown for illustration purposes) produced by the scanner 106. The scanner 106 receives returned or radiated energy and produces scanned images 108 that are sent to a threat detection system 110.

[0020] The threat detection system 110 processes the scanned images 108 to detect (preferably automatically) nuclear material or other contraband and communicate results of the detection to an operator station 114 via link 112. The threat detection system can be a stand alone system or form part of a larger security system. Link 112 can be a wired or wireless link such as a LAN, WAN, wireless network connection, radio link, optical link, or the like.

[0021] The energies 104a and 104b can include, for example, two or more different energy levels of x-ray energy. It will be appreciated that other types of electromagnetic energy can be used to scan the object 102. It will also be appreciated that although two energies (104a and 104b) are shown, more energy levels (e.g., four) can be used with an embodiment. Any type of scanner suitable for detecting contraband such as nuclear material and capable or producing an image (or array of values) may be used. The object being screened (or scanned) can include a cargo container, a truck, a tractor trailer, baggage, cargo, luggage, a vehicle, an air cargo container, and/or any object being transported that could potentially contain nuclear material or a portion of a threat or weapon system, or any object for which threat or contraband screening is contemplated or desired. The object being screened or scanned can also include a mail piece such as a letter, flat, package, parcel or the like. The radiographic images can be produced or generated at a shipping port, a border crossing, an airport, a truck terminal, or other facility or location where scanning or screening of objects using radiographic techniques may be desired. The images may be analyzed at the location where they are produced or may be provided to another location for analysis using a suitable communication method. Also, while the exemplary embodiments discussed herein are directed to detection of nuclear materials in cargo using threat scanning systems, it will be appreciated that the invention has application in other areas such as medical imaging and detection, material imaging for structural analysis or quality control, and the like. In general, the system described herein may be applied to any imaging context where detection of regions of interest having certain characteristics is desired.

[0022] FIG. 2 is a block diagram of an exemplary nuclear material detection system, showing greater detail. In particular, the threat detection system 110 includes a Z-analysis module 202, a segmentation processing module 204, a material context analysis module 206, and a threat decision arbitration module 208. Radiographic images 210 can be provided as input to the threat detection system 110. The output can include an indication of potential threat and/or false alarms 212.

[0023] In operation, the threat detection system 110 receives two or more radiographic images 210. The radiographic images, and associated data, can be provided in a proprietary format or in a standard format such as the N42 format, promulgated by American National Standards Institute (ANSI). Two or more images are typically used, and four images taken at four different energy levels can be particularly advantageous. The various energy levels provide different imaging characteristics. By using the different images for analysis, the advantages of each energy level can be realized, while attempting to reduce the disadvantages of each energy level. For example, while higher energy levels may provide better penetration through certain materials, the higher energy levels may saturate other materials. On the other hand, low energy levels may not provide as much penetration, but may also not have the saturation that accompanies higher energy levels. Thus, by using a combination of high and low energy levels, an embodiment may provide some of the benefits of each energy level and this may lead to a reduced false alarm rate and improved detection rate.

[0024] These images are typically first registered (or aligned) in order that subsequent analyses of the images are performed on corresponding portions of the images. Registration may be needed because the different images may be taken at different times with different imaging characteristics. Thus, registration may be needed for transforming the different sets of image data into one coordinate system. Registration may be done through a feature-based process or any other known or later developed registration method, such as area-based, transformation, search-based, spatial domain, frequency domain or the like. Two or more registration methods can be combined to register the images. In addition to the images, other data, such as a threat threshold, may also be provided as input. The registered images are provided as input to the Z-analysis processor (or routines) 202.

[0025] The Z-analysis module 202 determines an estimated atomic weight for the materials within the registered images. This determination can be performed at the pixel level, or at an object level including regions and layers to provide an enhanced analysis. If the determination of estimated atomic number is being performed at the object level, then object segmentation (described below) would be performed prior to Z-analysis. The Z-analysis module 202 can employ multiple algorithms to provide an enhance Z-analysis imaging capability. Once a Z-analysis has been performed, the pixel level Z-analysis is provided, along with the images, to the segmentation processing module 204.

[0026] The segmentation processing module 204 uses one or more image segmentation algorithms in order to identify objects in the image and report them to the Z-analysis module in the form of region of interest (ROI) coordinates. The threat objects being screened for by nuclear threat detection systems are typically dense and may appear solid in nature when analyzed. However, because various items in a cargo container may be layered between the scanner and the imaging device, overlapping regions of less dense material may appear as a denser and higher atomic number material. This poses a significant challenge for the segmentation processing module 204. In general there are four main approaches to image segmentation in order to separate an image into distinct objects that can be used, these are threshold, boundary, region-based and hybrid approaches. Region-based techniques include connected region analysis (CRA) and template region analysis (TRA) can be used. Another technique, independent component analysis (ICA) can be used. Also, an approach combining one or more of the above techniques may be used.

[0027] The Z-analysis module 202 can perform a subsequent Z-analysis at the object level on each of the ROIs returned by the segmentation module 204. The object level Z-analysis results (or effective Z values, Z.sub.eff) are then provided to the material context analysis module 206 for context and non-penetration analysis.

[0028] The material context analysis module 206 uses a-priori knowledge of the contents of the container and/or typical false alarms areas in order to analyze whether the ROIs received as input are potential threats or merely false alarms. The a-priori knowledge can be in the form of cargo manifests, expert system, historical knowledge, or the like. The suspect threat ROIs are provided as output from the material contest analysis module 206 to the Z-analysis module 202. Optionally, false alarm areas or other ROIs may be reported to the Z-analysis module 202 as well. The Z-analysis module can then provide the ROIs and associated confidence values to the threat decision arbitration module 208.

[0029] The threat decision arbitration module 208 can arbitrate between threat hypotheses produced by one threat detection system, or may arbitrate between results or hypotheses provided by multiple threat detection system of the same or different configuration. The threat decision arbitration module 208 uses expert-based rules to determine an optimal decision (given the inputs and rules) regarding the decision on potential threats and the confidences in those decisions. For example, artificial intelligence research has shown that arbitrating between the results of a plurality of different solutions or result sets may provide improved decision making ability for computerized systems under certain circumstances. Thus, the threat decision arbitration module 208 can accept input (e.g., potential threat ROIs and confidence values) from the Z-analysis module 202 and, optionally, from other internal or external systems or modules.

[0030] FIG. 3 is a block diagram of an exemplary nuclear material detection system showing interfaces between four processing modules. In particular, a Material Domain Imaging (MDI) module 302 has interfaces for receiving radiographic images 304 and also has interfaces to an Object Segmentation Recognition (OSR) module 306, a Material Context Analysis (MCA) module 312 and an Advanced Cognitive Arbitration (ACA) module 318. The interface for receiving the radiographic images 304 can include use of a proprietary or standard format, such as ANSI N42. In addition to the radiographic images, other data may be input to the MDI module 302, such as threat thresholds.

[0031] The interface between the MDI module 302 and the OSR module 306 includes input parameters 308 and output parameters 310, relative to the OSR module 306. The input parameters 308 include registered images and assigned Z-values. The registered images can be in gray scale and in an internal format. The output parameters 310 include region of interest (ROI) coordinates in an internal format.

[0032] The interface between the MDI module 302 and the MCA module 312 includes input parameters 314 and output parameters 316, relative to the MCA module 312. The input parameters 314 include assigned z-values and ROI coordinates (and can also include the hybrid images which contain the Zeff values and gray level data). Another input to the MCA module is configuration data. The configuration data can include container non-penetrable areas, a cargo manifest, and/or other encoded knowledge, data, or information that may be helpful in determining the context of ROIs. The output parameters 316 include non-penetrable regions and context suspicious regions. Optionally, false alarm regions may be output as well. The regions may be output as a set of coordinates.

[0033] The interface between the MDI module 302 and the ACA module 318 includes input parameters 320 and output results 322. The input parameters 320 can include one or more threat, warning, or false alarm ROIs and an associated confidence value for each. Another input to the ACA module 318 is expert rules that are used to determine an optimal output from the set of inputs received. The output results 322 include a decision array containing threats, warnings, and/or false alarms and associated confidence values for each. The output results 322 can include images and data in the ANSI N42 format. The output results can include a gray scale or colorized z-value image (where the gray value or color is based on the estimated atomic number determined by the MDI module 302), threat, warning or false alarm ROIs and associated confidences for each. Inputs to the ACA module 318 can also come from other radiographic systems, thus allowing the ACA module 318 to arbitrate between answers using data provided by various sources and/or vendors.

[0034] FIG. 4 is a flowchart of an exemplary method for detecting nuclear materials in a screened object. Processing begins at step 402 and continues to step 404.

[0035] In step 404, two or more radiographic images are received. These images can be in the format described above. Also, the images may be accompanied by other data, such as configuration parameters (information relating to the security system, scanning system, threat detection system or object being scanned) and/or a threat threshold. Once the radiographic images are received, control continues to step 406.

[0036] In step 406, a Z-analysis as described above is performed. The Z-analysis results in a Z-value map or array of estimated material atomic numbers that corresponds to two or more of the radiographic images. Control continues to step 408.

[0037] In step 408, object segmentation is performed on the radiographic images. As mentioned above, it may be desirable to perform object segmentation prior to Z-analysis, in which case step 408 may be performed before step 406. Also, a second Z-analysis can be performed after the object segmentation and step 406 could be repeated after step 408 using the object regions of interest (ROIs) identified by the segmentation process. Control continues to step 410. The segmentation process can use as few as one image; however the Z-analysis may require at least two images using different energies.

[0038] In step 410, the material context of any regions of interest is analyzed to help identify both false alarm areas and potential threat areas. Control continues to step 412.

[0039] In step 412, possible or potential threat ROIs are arbitrated using a set of expert rules. Control continues to step 414.

[0040] In step 414, potential threats, warnings, and/or false alarm ROIs are provided as output. Control continues to step 416 where the method ends.

[0041] It will be appreciated that steps 404-414 may be repeated in whole or in part in order to accomplish a contemplated nuclear material detection task. Further, it should be appreciated that the steps mentioned above may be performed on a single or distributed processor. Also, the processes, modules, and sub-modules described in the various figures of the embodiments above may be distributed across multiple computers or systems or may be co-located in a single processor or system.

[0042] The modules, processors or systems described above can be implemented as a programmed general purpose computer, an electronic device programmed with microcode, a hard-wired analog logic circuit, software stored on a computer-readable medium or signal, a programmed kiosk, an optical computing device, a GUI on a display, a networked system of electronic and/or optical devices, a special purpose computing device, an integrated circuit device, a semiconductor chip, and a software module or object stored on a computer-readable medium or signal, for example.

[0043] Embodiments of the method and system for nuclear material detection (or their sub-components), may be implemented on a general-purpose computer, a special-purpose computer, a programmed microprocessor or microcontroller and peripheral integrated circuit element, an ASIC or other integrated circuit, a digital signal processor, a hardwired electronic or logic circuit such as a discrete element circuit, a programmed logic circuit such as a PLD, PLA, FPGA, PAL, or the like. In general, any process capable of implementing the functions or steps described herein can be used to implement embodiments of the method, system, or a computer program product (software program) for nuclear material detection.

[0044] Furthermore, embodiments of the disclosed method, system, and computer program product for nuclear material detection may be readily implemented, fully or partially, in software using, for example, object or object-oriented software development environments that provide portable source code that can be used on a variety of computer platforms. Alternatively, embodiments of the disclosed method, system, and computer program product for nuclear material detection can be implemented partially or fully in hardware using, for example, standard logic circuits or a VLSI design. Other hardware or software can be used to implement embodiments depending on the speed and/or efficiency requirements of the systems, the particular function, and/or particular software or hardware system, microprocessor, or microcomputer being utilized. Embodiments of the method, system, and computer program product for nuclear material detection can be implemented in hardware and/or software using any known or later developed systems or structures, devices and/or software by those of ordinary skill in the applicable art from the function description provided herein and with a general basic knowledge of the computer, image processing, radiographic, and/or threat detection arts.

[0045] Moreover, embodiments of the disclosed method, system, and computer program product for nuclear material detection can be implemented in software executed on a programmed general purpose computer, a special purpose computer, a microprocessor, or the like. Also, the method for nuclear material detection of this invention can be implemented as a program embedded on a personal computer such as a JAVA.RTM. or CGI script, as a resource residing on a server or image processing workstation, as a routine embedded in a dedicated processing system, or the like. The method and system can also be implemented by physically incorporating the method for nuclear material detection into a software and/or hardware system, such as the hardware and software systems of multi-energy radiographic cargo inspection systems.

[0046] It is, therefore, apparent that there is provided, in accordance with the present invention, a method, computer system, and computer software program for nuclear material detection. While this invention has been described in conjunction with a number of embodiments, it is evident that many alternatives, modifications and variations would be or are apparent to those of ordinary skill in the applicable arts. Accordingly, Applicant intends to embrace all such alternatives, modifications, equivalents and variations that are within the spirit and scope of this invention.

Claims

1. A method for automatically detecting nuclear material using radiographic images of a cargo container, the method comprising: receiving a plurality of radiographic images of the cargo container; aligning the plurality of images with respect to each other to produce registered images; segmenting the registered images using the atomic number information in order to locate one or more regions of interest within the registered images; estimating atomic number information for each of a predetermined number of portions of the registered images; assigning a threat level and a confidence value to regions of interest identified as a potential threat; evaluating the regions of interest identified as potential threats using material context information and adjusting, based on the evaluation, the threat level values and confidences of the regions of interest identified as potential threats; and providing the regions of interest and adjusted threat level and confidence values as output to an operator station.

2. The method of claim 1, further comprising scanning the cargo container with an imaging device to produce the plurality of radiographic images.

3. The method of claim 1, wherein the imaging device includes a linear accelerator and detector combination.

4. The method of claim 1, wherein the each of the radiographic images is produced using a different energy level.

5. The method of claim 1, further comprising producing the radiographic images at a shipping port.

6. The method of claim 5, wherein the evaluating is performed at a geographic location remote from the port.

7. The method of claim 1, wherein the step of estimating atomic number information includes generating a gray scale image having gray values that correspond to an estimated atomic number of a material being scanned.

8. The method of claim 6, wherein the operator station is located at the port.

9. The method of claim 1, wherein the material context information includes a cargo manifest.

10. A system for automatically detecting nuclear material in a radiographic image of a cargo conveyance, the system comprising: a material domain imaging module adapted to receive a plurality of radiographic images and to determine an atomic number for each pixel in a combined atomic number image, the atomic number based on an analysis of a pixel in each of the radiographic images; an object segment recognition module coupled to the material domain imaging module and adapted to perform an image segmentation operation on the combined atomic number image and to identify and return one or more object regions to the material domain imaging module, whereby the material domain imaging module can perform an object level atomic number analysis using a region of the combined atomic number image defined by the one or more object regions; a material context analysis module coupled to the material domain imaging module and adapted to receive the combined atomic number image, having been segmented into the one or more object regions, and to generate a hypothesis as to whether an object region that has been assigned a high atomic number is in a suspicious location of the image, the hypothesis generation based on a computer-usable representation of expert knowledge and historical data, the material context analysis module also being adapted to provide the hypothesis as output to another module; and an advanced cognitive arbitration module coupled to the material domain imaging module and adapted to receive a plurality of threat hypotheses as inputs, each having an associated confidence value, the advanced cognitive arbitration module adapted to rank the inputs, assign a confidence value to each ranked input and provide a ranked list of threats, with associated confidence values, as output.

11. The system of claim 10, wherein the cargo conveyance includes a cargo container.

12. The system of claim 11, wherein the cargo conveyance further includes a truck and a trailer onto which the cargo container has been loaded.

13. The system of claim 10, wherein the plurality of radiographic images includes four images each produced using a different energy level.

14. The system of claim 10, wherein the material domain imaging module determines an atomic number estimate for each of the object regions identified by the object segment recognition module.

15. A threat detection system comprising: means for determining estimated material atomic number values based on two or more radiographic images produced using a plurality of energy levels and generating an estimated atomic number gray level image, the gray level being based on the estimated material atomic number values; means for segmenting the gray level image to define one or more regions of interest; means for analyzing a threat level of each region of interest using material context information; and means for arbitrating among multiple potential threat results to determine a final threat decision array for output.

16. The threat detection system of claim 15, wherein the plurality of energy levels includes at least four energy levels.

17. The threat detection system of claim 15, wherein at least one of the multiple threat results are received from another threat detection system different from said threat detection system.

18. The threat detection system of claim 15, wherein the threat decision array includes a confidence value corresponding to each potential threat.

19. The threat detection system of claim 15, wherein the means for determining estimated material atomic number values further includes determining an estimated material atomic value for each region of interest.

20. The threat detection system of claim 15, wherein the material context information includes a cargo manifest listing the contents of a cargo conveyance being screened for threats.