Systems and methods for detecting threats and contraband in cargo
View Patent ↗The present specification discloses systems and methods for identifying and reporting contents of a tanker, container or vehicle. Programmatic tools are provided to assist an operator in analyzing contents of a tanker, container or vehicle. Manifest data is automatically imported into the system for each shipment, thereby helping security personnel to quickly determine container contents. In case of a mismatch between container contents shown by manifest data and the contents as ascertained from the scanning system, the container or vehicle may be withheld for further inspection.
1. A method for processing a radiographic image to identify whether at least one firearm is present in an object, the method comprising:
receiving the radiographic image of the object;
applying a neural network to the radiographic image to minimize errors in an association of portions of the radiographic image with one or more predetermined features; and
based on the association with minimized errors, generating a visual demarcation of an area in the radiographic image to identify a presence of the at least one firearm in the object.
2. The method of claim 1 , wherein the neural network comprises a plurality of interconnected layers defined by variables associated with the one or more predetermined features, wherein the one or more predetermined features are indicative of a firearm.
3. The method of claim 1 , further comprising detecting the at least one firearm in more than one physical orientation.
4. The method of claim 1 , further comprising generating reference images wherein the reference images depict firearms in a plurality of physical orientations.
5. The method of claim 4 , further comprising identifying the presence of the at least one firearm even if the at least one firearm does not exactly match a firearm in one of the reference images.
6. The method of claim 1 , wherein the neural network is a deep belief network.
7. The method of claim 1 , wherein the radiographic image comprises attenuation data.
8. The method of claim 1 , further comprising building orientation invariant descriptors of features in radiographic images.
9. The method of claim 1 , further comprising applying one or more classifier routines to the one or more features in the radiographic image.
10. The method of claim 9 , wherein the one or more features in the radiographic image comprises at least one effective density, attenuation, geometric properties, size, area, or aspect ratio.
11. A computer readable non-transitory medium comprising a plurality of executable programmatic instructions executed by a processor for implementing a process for processing a radiographic image to identify whether at least one firearm is present in an object, said plurality of executable programmatic instructions comprising:
programmatic instructions, stored in said computer readable non-transitory medium, for receiving the radiographic image of the object;
programmatic instructions, stored in said computer readable non-transitory medium, for applying a neural network to the radiographic image to minimize errors in an association of portions of the radiographic image with one or more predetermined features; and
programmatic instructions, stored in said computer readable non-transitory medium, for generating a visual demarcation of an area in the radiographic image to identify a presence of the at least one firearm in the object based on the association with minimized errors.
12. The computer readable non-transitory medium of claim 11 , wherein the neural network comprises a plurality of interconnected layers defined by variables associated with the one or more predetermined features, wherein the one or more predetermined features are indicative of a firearm.
13. The computer readable non-transitory medium of claim 11 , further comprising programmatic instructions, stored in said computer readable non-transitory medium, for detecting the at least one firearm in more than one physical orientation.
14. The computer readable non-transitory medium of claim 11 , further comprising programmatic instructions, stored in said computer readable non-transitory medium, for generating reference images wherein the reference images depict firearms in a plurality of physical orientations.
15. The computer readable non-transitory medium of claim 14 , further comprising programmatic instructions, stored in said computer readable non-transitory medium, for identifying the presence of the at least one firearm even if the at least one firearm does not exactly match a firearm in one of the reference images.
16. The computer readable non-transitory medium of claim 11 , wherein the neural network is a deep belief network.
17. The computer readable non-transitory medium of claim 11 , wherein the radiographic image comprises attenuation data.
18. The computer readable non-transitory medium of claim 11 , further comprising programmatic instructions, stored in said computer readable non-transitory medium, for building orientation invariant descriptors of the one or more features in radiographic images.
19. The computer readable non-transitory medium of claim 11 , further comprising programmatic instructions, stored in said computer readable non-transitory medium, for applying one or more classifier routines to the one or more features in the radiographic image.
20. The computer readable non-transitory medium of claim 19 , wherein the one or more features in the received radiographic image comprises at least one effective density, attenuation, geometric properties, size, area, or aspect ratio.