Detection of a concealed object

• US 7,834,802 B2
• Filed: 04/28/2008
• Issued: 11/16/2010
• Est. Priority Date: 03/16/2001
• Status: Expired due to Term

• Current Assignee: Battelle Memorial Institute
• Original Assignee: Battelle Memorial Institute
• Inventors: Hall, Thomas E., Keller, Paul E., McMakin, Douglas L.
• Primary Examiner(s): ALSOMIRI, ISAM A
• Application Number:
  US12/111,023
• Publication Number:
  US 20090140907A1
• Time in Patent Office: 932 Days
• Field of Search: 342/22, 342/27, 342/42, 342/44, 342/179
• US Class Current: 342/22
• CPC Class Codes: G01N 21/3563 for analysing solids; Preparation of samples therefor
• CPC Class Codes: G01N 21/3581 using far infrared light; using Terahertz radiation
• CPC Class Codes: G01S 13/34 using transmission of continuous, frequency-modulated waves while heterodyning the received signal, or a signal derived therefrom, with a locally-generated signal related to the contemporaneously transmitted signal
• CPC Class Codes: G01S 13/887 for detection of concealed objects, e.g. contraband or weapons
• CPC Class Codes: G01S 13/89 for mapping or imaging
• CPC Class Codes: G01S 7/20 Stereoscopic displays; Three-dimensional displays; Pseudo-three-dimensional displays
• CPC Class Codes: G01S 7/35 Details of non-pulse systems
• CPC Class Codes: G01S 7/41 using analysis of echo signal for target characterisation; Target signature; Target cross-section
• CPC Class Codes: G01S 7/417 involving the use of neural networks

Abstract

Disclosed are systems, methods, devices, and apparatus to determine if a clothed individual is carrying a suspicious, concealed object. This determination includes establishing data corresponding to an image of the individual through interrogation with electromagnetic radiation in the 200 MHz to 1 THz range. In one form, image data corresponding to intensity of reflected radiation and differential depth of the reflecting surface is received and processed to detect the suspicious, concealed object.

Claims

1. A method, comprising:

detecting electromagnetic radiation returned from a concealed surface associated with a person;

generating image data from the detected electromagnetic radiation;

segmenting the image data into image portions;

performing a Fourier transform on each of the image portions;

identifying the presence of a man-made object in one or more of the image portions based on spatial frequencies in the Fourier-transformed image portions, wherein the identifying the presence of the man-made object comprises applying at least one extraction filter to the image portion, thereby extracting features from the image portions, wherein the at least one extraction filter is a rotationally invariant extraction filter or a scale invariant extraction filter; and

communicating the presence of the man-made object.

2. The method of claim 1, wherein the identifying the presence of the man-made object includes identifying image portions having a higher percentage of spatial frequencies in a spatial frequency range than other image portions.

3. The method of claim 1, wherein the at least one extraction filter comprises a ring-wedge-based feature extractor, or a ring-based feature extractor.

4. The method of claim 1, wherein the identifying the presence of the man-made object further comprises adaptively processing the extracted features using a neural network.

5. The method of claim 4, wherein the neural network is a multilayer perceptron neural network.

6. The method of claim 1, wherein the image portions include overlapping image portions, wherein the identifying the presence of the man-made object comprises adaptively processing the overlapping image portions, and wherein the method further comprises comparing results from adaptively processing the overlapping image portions.

7. The method of claim 1, further comprising interrogating the concealed surface associated with the person with electromagnetic radiation at one or more frequencies in a range of about 200 MHz to about 1 THz.

8. A system, comprising:

an array operable to interrogate a person with electromagnetic radiation at one or more frequencies in a frequency range; and

a processing subsystem coupled to the array, the processing subsystem being operable to transform image data including intensity and depth data produced from returned electromagnetic radiation into one or more spatial frequency representations, to apply at least one extraction filter to the spatial frequency representations, and to adaptively process extracted features output from the one or more extraction filters so as to determine whether one or more objects being concealed by the person are man-made objects, wherein the at least one extraction filter is a rotationally invariant extraction filter or a scale invariant extraction filter.

9. The system of claim 8, wherein the processing subsystem implements a neural network configured to adaptively process the extracted features.

10. The system of claim 8, further comprising a display device configured to provide at least one image if the processing subsystem determines that the one or more objects being concealed by the person are man-made objects.

11. The system of claim 8, wherein the array is a first array and is provided in a first panel, and wherein the system further comprises a second array in a second panel coupled to the processing subsystem.

12. The system of claim 8, wherein the processing subsystem is further operable to generate image data corresponding to a number of cylindrical images of the person.

13. The system of claim 8, wherein the frequency range is about 200 MHz to about 1 THz.

14. A method, comprising:

detecting electromagnetic radiation returned from a concealed surface associated with a person;

generating image data from the detected electromagnetic radiation;

segmenting the image data into image portions;

transforming the image portions in spatial frequency representations;

extracting features from the spatial frequency representations based on spatial frequencies of the features, wherein the extracting the features comprise applying at least one extraction filter, that is a rotationally invariant extraction filter or a scale invariant extraction filter to the spatial frequency representations; and

identifying the presence of a man-made object in one or more of the image portions based on the extracted features.

15. The method of claim 14, wherein the at least one extraction filter is a ring-wedge-based feature extractor, a sector-based feature extractor, or a ring-based feature extractor.

16. The method of claim 14, wherein the identifying the presence of the man-made object further comprises inputting the extracted features into a neural network.

17. The method of claim 16, wherein the neutral network is of a multilayer perceptron form.

18. The method of claim 14, further comprising interrogating the concealed surface associated with the person with electromagnetic radiation at one or more frequencies in a range of about 200 MHz to about 1 THz to provide the corresponding electromagnetic radiation returned from the concealed surface.

19. The method of claim 14, wherein the transforming the image portions comprises performing a Fourier transform of the image portions.