Task-Based Imaging Systems

• US 20100272327A1
• Filed: 07/02/2010
• Published: 10/28/2010
• Est. Priority Date: 12/01/2003
• Status: Active Grant

• Current Assignee: Edward R. Dowski, Jr., Gregory E. Johnson., Paulo E.X. Silveira, Ramkumar Narayanswamy, Robert H. Cormack
• Original Assignee: Edward R. Dowski, Jr., Gregory E. Johnson., Paulo E.X. Silveira, Ramkumar Narayanswamy, Robert H. Cormack
• Inventors: Narayanswamy, Ramkumar, Johnson, Gregory E., Dowski, Edward R. JR., Silveira, Paulo E.X., Cormack, Robert H.
• Granted Patent: US 8,144,208 B2
• Time in Patent Office: Days
• US Class Current: 382/117
• CPC Class Codes: G02B 26/06 for controlling the phase of light G02B26/08 takes precedence , measuring optical phase difference G01J9/00
• CPC Class Codes: G02B 27/0075 with means for altering, e.g. increasing, the depth of field or depth of focus
• CPC Class Codes: G06V 40/18 Eye characteristics, e.g. of the iris

Abstract

A method for generating an output image of a scene is disclosed. A detector of a task-based imaging system includes a plurality of pixels, and the scene includes at least one object located at a given object distance within a range of object distances between the object and the imaging system. The method includes capturing a high resolution image of the scene, converting the high resolution image into an image spectrum of the scene, determining a defocused optical transfer function (OTF) of the imaging system over the range of object distances and determining a pixel modulation transfer function (MTF) over the plurality of pixels. The method also includes multiplying the image spectrum with the OTF and the MTF to generate a modified image spectrum of the scene, converting the modified image spectrum into a modified image of the scene, and generating the output image from the modified image.

Claims

1. A method for generating an output image of a scene captured by a detector of a task-based imaging system, the detector including a plurality of pixels and the scene including at least one object located at a given object distance within a range of object distances, which object distance is defined as a distance between the object and the task-based imaging system, the method comprising:

capturing a high resolution image of the scene over a range of spatial frequencies;

converting the high resolution image into an image spectrum of the scene;

determining a defocused optical transfer function (OTF) of the task-based imaging system over the range of object distances;

determining a pixel modulation transfer function (MTF) over the plurality of pixels of the detector;

multiplying the image spectrum with the OTF and the MTF to generate a modified image spectrum of the scene;

converting the modified image spectrum into a modified image of the scene;

generating the output image from the modified image.

2. The method of claim 1, wherein generating the output image includes forming a resized image from the modified image, and forming the output image from the resized image.

3. The method of claim 2, wherein forming the resized image includes down-sampling the modified image to a final resolution, in accordance with the given object distance.

4. The method of claim 3, wherein down-sampling includes resizing the modified image without low-pass filtering.

5. The method of claim 3, wherein down-sampling comprises:

processing the modified image for a given down-sampling origin and a down-sampling period; and

generating multiple aliased versions of the resized image by varying the down-sampling origin within the down-sampling period.

6. The method of claim 2, wherein the detector has shot noise and read noise characteristics, and wherein forming the output image comprises adding at least one of the shot noise and read noise characteristics to the resized image.

7. The method of claim 2, wherein the resized image includes at least one boundary line, the method further comprising padding the resized image by replicating the boundary line to generate a padded image with a desired image size.

8. The method of claim 1, further comprising filtering the output image with a filter kernel to generate a filtered image.

9. The method of claim 1, further comprising:

from the output image, calculating a signal-to-noise ratio (SNR) over the scene;

modifying the defocused OTF of the imaging system such that the SNR in the output image is larger than that calculated from the output image without modifying the defocused OTF.

10. The method of claim 9, further comprising:

averaging the SNR over a range of directions over the scene to generate an average SNR; and

altering the defocused OTF such that the average SNR in the output image is larger than that calculated from the output image without altering the defocused OTF.

11. The method of claim 10, wherein averaging the SNR comprises calculating a weighted average SNR over the scene.

12. The method of claim 11, wherein calculating the weighted average SNR includes varying a weighting factor of the SNR in accordance with at least a selected one of a reduction in signal strength and a reduction in incoherent imaging system modulation as a function of the given object distance.

13. The method of claim 11, further comprising setting a focus of the task-based imaging system at a position corresponding to that at which an SNR calculated for smallest and largest object distance values in the range of object distances are equal.

14. The method of claim 10, wherein altering comprises modifying the task-based imaging system to effect a pupil function of a form:

15. The method of claim 1, further comprising post-processing the output image for performing at least one task selected from the group consisting of biometric iris recognition, biometric face recognition, biometric recognition for access control, biometric recognition for threat identification, barcode reading, imaging for quality control in an assembly line, optical character recognition, biological imaging and automotive imaging for object detection.

16. A method for optimizing a task-based imaging system for obtaining data of a scene for use in a task over a range of object distances, the scene including at least one object located at a given object distance within the range of object distances, which object distance is defined as a distance between the object and the task-based imaging system, the method comprising:

1) determining a target signal-to-noise ratio (SNR) of the task-based imaging system;

2) specifying an initial set of pupil function parameters and a merit function;

3) generating a new set of pupil function parameters based on the merit function;

4) calculating SNR over the range of object distances;

5) comparing the SNR to a target SNR; and

6) repeating steps

     2) through

     5) until the SNR is at least equal in value to the target SNR.

17. The method of claim 16,

wherein determining the target SNR comprises determining a target polar-SNR, and

wherein calculating comprises calculating a polar-SNR over the range of object distances.