Data-efficient and self adapting imaging spectrometry method and an apparatus thereof

• US 6,920,244 B2
• Filed: 10/05/2001
• Issued: 07/19/2005
• Est. Priority Date: 10/06/2000
• Status: Active Grant

• Current Assignee: Rochester Institute of Technology
• Original Assignee: Rochester Institute of Technology
• Inventors: Ohta, Noboru, Rosen, Mitchell
• Primary Examiner(s): Wu, Jingge
• Application Number:
  US09/972,223
• Publication Number:
  US 20020090144A1
• Time in Patent Office: 1,383 Days
• Field of Search: 382/110, 382162-167, 382254-256, 382270-272, 25033001-33014
• US Class Current: 382/162
• CPC Class Codes: G01J 3/28 Investigating the spectrum using colour filters G01J3/51
• CPC Class Codes: G01J 3/2823 Imaging spectrometer
• CPC Class Codes: H04B 2203/545 Audio/video application, e.g. interphone
• CPC Class Codes: H04B 3/54 Systems for transmission via power distribution lines

Abstract

A method for spectral imaging includes capturing high spectral resolution data of at least a portion of an image using a plurality of channels, determining a first set of channels from the plurality of channels which can reconstruct spectra of the portion of the image to within a first error tolerance from the captured high spectral resolution data, and capturing estimated spectral resolution data of the image using the first set of channels.

Claims

1. A method for spectral imaging, the method comprising:

capturing high spectral resolution data of at least a first portion of a first scene using a first plurality of channels;

determining a first set of channels from a second plurality of channels which can reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data, wherein the first set of channels from the second plurality of channels comprises a smallest number of the plurality of channels which can be used to reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data, wherein the determining further comprises performing an iterative optimization to determine a first set of channels from a second plurality of channels which can reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data; and

capturing pixel data of at least a second portion of at least the first scene using the first set of channels.

2. The method as set forth in claim 1 wherein the first set of channels from the second plurality of channels comprises a first stored number of the plurality of channels or a subset of the first stored number of the second plurality of channels which can be used to reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data.

3. The method as set forth in claim 1 further comprising determining a first transform from the first set of channels which can reconstruct spectra.

4. The method as set forth in claim 3 further comprising reproducing the image spectra using the first transform and the captured pixel data from the at least a second portion of at least the first scene.

5. The method as set forth in claim 3 further comprising storing the first transform.

13. The method as set forth in claim 1 further comprising storing the captured pixel data from the first set of channels from the second portion of the first scene.

17. The method as set forth in claim 1 wherein the first error criterion requires that the reconstructed spectra of the first portion of the first scene be within a first error tolerance from the captured high spectral resolution data or that the reconstructed spectra of the first portion of the first scene be associated with a minimum value for a predetermined metric when compared with the captured high spectral resolution data.

20. The method as set forth in claim 1 wherein the number of the first plurality of channels and the second plurality of channels are identical.

6. A method for spectral imaging, the method comprising:

capturing high spectral resolution data of at least a first portion of a first scene using a first plurality of channels;

determining a first set of channels from a second plurality of channels which can reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data;

capturing pixel data of at least a second portion of at least the first scene using the first set of channels;

determining a first transform from the first set of channels which can reconstruct spectra;

at least once capturing high spectral resolution data of at least a third portion of a second scene;

capturing pixel data of the at least a third portion of a second scene using the first set of channels;

reconstructing estimated spectra by applying the first transform to the captured pixel data;

comparing the high spectral resolution data with the reconstructed estimated spectra to determine an intermittent error;

determining a second set of channels from the second plurality of channels which can reconstruct spectra of the third portion of the second scene to satisfy the first error criterion when compared with the captured high spectral resolution data if the intermittent error does not satisfy a second error criterion; and

capturing pixel data of at least a fourth portion of any remaining portion of at least the second scene using the second set of channels.

7. The method as set forth in claim 6 wherein the first and second error criteria are the same.

8. The method as set forth in claim 6 further comprising capturing high spectral resolution data of the third portion of the second scene using the first plurality of channels.

9. The method as set forth in claim 6 wherein the second set of channels from the second plurality of channels comprises a first stored number of the plurality of channels or a subset of the first stored number of the plurality of channels which can be used to reconstruct spectra of the third portion of the second scene to satisfy a second error criterion when compared with the captured high spectral resolution data.

10. The method as set forth in claim 6 further comprising determining a second transform from the second set of channels which can reconstruct spectra.

11. The method as set forth in claim 10 further comprising reproducing the image spectra using the first and second transforms from the first and second sets of channels and the captured pixel data from the at least a second and fourth portions of at least the first and second scenes.

12. The method as set forth in claim 10 further comprising storing the second transform.

14. The method as set forth in claim 6 further comprising storing the captured pixel data from the second portion of the first scene such that the captured pixel data from the second portion of the first scene remains associated with the first transform.

15. The method as set forth in claim 6 further comprising storing the captured pixel data from the second set of channels from the at least a fourth portion of at least the second scene.

16. The method as set forth in claim 14 further comprising storing the captured pixel data from the fourth portion of the second scene such that the captured pixel data from the fourth portion remains associated with the second transform.

18. The method as set forth in claim 7 wherein the second error criterion requires that the reconstructed spectra of the third portion of the second scene be within a second error tolerance from the captured high spectral resolution data or that the reconstructed spectra of the third portion of the second scene be associated with a minimum value for a predetermined metric when compared with the captured high spectral resolution data.

19. A method for spectral imaging, the method comprising:

capturing high spectral resolution data of at least a first portion of a first scene using a first plurality of channels;

determining a first set of channels from a second plurality of channels which can reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data;

capturing pixel data of at least a second portion of at least the first scene using the first set of channels;

determining which aspects of the captured estimated spectral resolution data are signatures of original color levels and which aspects of the captured estimated spectral resolution data are signatures of deterioration artifacts based on stored image information; and

correcting the aspects of the captured estimated spectral resolution data which are signatures of deterioration artifacts.

21. A system for spectral imaging, the system comprising:

an imaging system that captures high spectral resolution data of at least a portion of a first scene using a first plurality of channels; and

a spectral processing system that determines a first set of channels from a second plurality of channels which can reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data, wherein the first set of channels from the plurality of channels comprises a smallest number of channels which can be used to reconstruct spectra of the first portion of the scene to satisfy a first error criterion when compared with the captured high spectral resolution data, wherein the imaging system captures pixel data of at least a second portion of at least the first scene using the first set of channels, wherein the spectral processing system performs an iterative optimization to determine a first set of channels from a second plurality of channels which can reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data.

22. The system as set forth in claim 21 wherein the first set of channels from the first plurality of channels comprises a first stored number of the plurality of channels or a subset of the first stored number of the plurality of channels which can be used to reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data.

23. The system as set forth in claim 21 wherein the spectral processing system determines a transform from the first set of channels and the imaging system reproduces the image using the transform and the captured pixel data.

24. The system as set forth in claim 21 wherein the spectral processing system stores the transform from the first set of channels.

30. The system as set forth in claim 21 wherein the number of the first plurality of channels and the second plurality of channels are identical.

38. The system as set forth in claim 21 wherein the imaging system further comprises:

a first imaging system that captures the high spectral resolution data of at least the portion of the first scene using the plurality of channels; and

a second imaging system captures the pixel data of the scene using the first set of channels.

25. A system for spectral imaging, the system comprising:

an imaging system that captures high spectral resolution data of at least a portion of a first scene using a first plurality of channels; and

a spectral processing system that determines a first set of channels from a second plurality of channels which can reconstruct spectra of the first portion of the first scene to satisfy a first error criterion when compared with the captured high spectral resolution data, wherein the imaging system captures pixel data of at least a second portion of at least the first scene using the first set of channels;

wherein the spectral processing system compares at least once the high spectral resolution data for a second portion of a second scene against the estimated spectral resolution data captured using the first set of channels for the second portion of the second scene to determine an intermittent error, the spectral processing system determines a second set of channels from the second plurality of channels which can reconstruct spectra of the second portion of the second scene to satisfy the first error criterion when compared to the captured high spectral resolution data if the intermittent error is greater then a second error tolerance, and the imaging system captures pixel data of at least a portion of any remaining portion of at least the second scene using the second set of channels.

26. The system as set forth in claim 25 wherein the imaging system captures high spectral resolution data of the second portion of the image using the first plurality of channels.

27. The system as set forth in claim 25 wherein the spectral processing system determines a second transform from the second set of channels and the imaging system reproduces the image using the first and second transforms and the pixel data captured using the first and second sets of channels.

28. The system as set forth in claim 25 wherein the spectral processing system determines a transform from the first set of channels and stores the transform and the pixel data captured using the first and second sets of channels.

29. The system as set forth in claim 25 wherein the spectral processing system determines which aspects of the captured estimated spectral resolution data are signatures of original color levels and which aspects of the captured estimated spectral resolution data are signatures of deterioration artifacts based on stored image information and wherein the spectral processing system corrects the aspects of the captured estimated spectral resolution data which are signatures of deterioration artifacts.

31. A method for spectral imaging, the method comprising:

capturing a first high spectral resolution data of at least a first portion of a first scene using a plurality of channels;

capturing pixel data of at least a second portion of a first scene using a first set of channels from the first plurality of channels;

determining a first transform based on the first set of channels and the first high spectral resolution data;

generating an image of the first scene using the transform and the captured pixel data;

capturing high spectral resolution data of at least a second portion of a second scene using a plurality of channels;

applying the first transform to the pixel data from the first set of channels to the second portion of the second scene to produce spectral estimates;

compare the spectral estimates to the high spectral resolution data to determine an intermittent error; and

determining a second transform based on the first set of channels and the second high spectral resolution data if the intermittent error is greater than a first error tolerance.

32. The method as set forth in claim 31 further comprising recording the generated image on a media.

33. The method as set forth in claim 31 further comprising storing the generated image.

34. A system for spectral imaging, the system comprising:

a first imaging sub-system that captures a first high spectral resolution data of at least a first portion of a first scene using a plurality of channels;

a second imaging sub-system that captures pixel data of at least a second portion of a first scene using a first set of channels from a plurality of channels; and

a spectral processing system that determines a first transform based on the first set of channels and the first high spectral resolution data generates the image of the first scene using the transform and the captured pixel data;

wherein the first imaging sub-system captures high spectral resolution data of at least a portion of a second scene using a plurality of channels and wherein the spectral imaging system applies the first transform to the pixel data from the first set of channels of the at least a portion of the second scene producing spectral estimates, compares the spectral estimates to the high spectral resolution data to determine an intermittent error and determines a second transform based on the first set of channels and the second high spectral resolution data if the intermittent error is greater than a first error tolerance.

35. The system as set forth in claim 34 wherein a recording system records the generated image on a media.

36. The system as set forth in claim 34 wherein the spectral processing system stores the generated image.

37. The system as set forth in claim 34 wherein the first and second imaging sub-system are the same imaging system.