SCENE BASED NON-UNIFORMITY CORRECTION FOR INFRARED DETECTOR ARRAYS

• US 20120091340A1
• Filed: 10/19/2010
• Published: 04/19/2012
• Est. Priority Date: 10/19/2010
• Status: Active Grant

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: YOUNG, Ching-Ju J.
• Granted Patent: US 8,203,116 B2
• Time in Patent Office: Days
• US Class Current: 250/332
• CPC Class Codes: H04N 25/671 for non-uniformity detection or correction
• CPC Class Codes: H04N 25/674 based on the scene itself, e.g. defocusing
• CPC Class Codes: H04N 25/683 by defect estimation performed on the scene signal, e.g. real time or on the fly detection
• CPC Class Codes: H04N 5/33 Transforming infrared radiation cameras or camera modules for generating image signals from infrared radiation H04N23/20; circuitry of SSIS for transforming infrared radiation into image signals H04N25/20

Abstract

In various embodiments, a method and system for compensating non-uniformities among detector elements of a detector array, without the use of dither mirrors or requirement of scene motion for non-uniformity correction achieved by computing scene spatial gradient and temporal gradient of image frames of the scene captured by the detector array at different times, and utilizing both the scene spatial and temporal gradients in detailed local gradient processing. Such local gradient processing may include computing masks to preserve spatial scene details, while eliminating scene noise (e.g., fixed pattern noise) from the captured image frames and correcting non-uniformity among detector elements.

Claims

1. A ditherless method of compensating non-uniformities among detector elements of a detector array that captures successive image frames of a scene, where each detector element of the array has at least one neighbor, the method comprising:

(i) receiving a current image frame from the detector array and a previous image frame from an image frame memory, wherein the current and previous frames are image frames captured by the detector array at different times and which are represented as respective matrices each containing image pixel values representing output signals from the detector array;

(ii) applying, at a correction unit, a plurality of predetermined correction terms to the current frame and the previous frame to generate a corrected current frame and a corrected previous frame;

(iii) generating, at a delta-frame unit, a delta frame using the current frame and the previous frame;

(iv) computing, at a gradient-calculation unit, a previous-frame gradient matrix of the corrected previous frame in a first direction, a current-frame gradient matrix of the corrected current frame in a second direction, and a first delta-frame gradient matrix in the first direction and a second delta-frame gradient matrix in the second direction of the delta frame, wherein the second direction is opposite to the first direction;

(v) computing, at a gradient-processing unit, a first mask using the previous-frame gradient matrix and the first delta-frame gradient matrix, and a second mask using the current-frame gradient matrix and the second delta-frame gradient matrix, wherein the first mask preserves scene-gradient information of a scene associated with the corrected previous frame in the first direction, and the second mask preserves scene-gradient information associated with the corrected current frame in the second direction;

(vi) computing, at a delta-correction unit, a plurality of delta-correction values as a function of the previous-frame gradient matrix, the current-frame gradient matrix, and the first and second masks, wherein each of the plurality of delta-correction values is associated with a different one of the detector elements of the array; and

(vii) updating, at an update unit, each of the plurality of predetermined correction terms with a corresponding delta-correction value.

2. The method of claim 1, wherein the detector array comprises a one-dimensional linear array.

3. The method of claim 1, further comprising using local scene data to compensate for non-uniformities for only a portion of detector elements of the detector array.

4. The method of claim 1, wherein the detector array comprises a two-dimensional planar array.

5. The method of claim 4 further comprising repeating steps (iv) through (vii) for third and fourth directions perpendicular to the first and second directions, respectively.

6. The method of claim 1, wherein computing the first mask includes comparing values of the previous-frame gradient matrix and values of the first delta-frame gradient matrix with a predetermined threshold value.

7. The method of claim 6, wherein computing the second mask includes comparing values of the current-frame gradient matrix and values of the second delta-frame gradient matrix with the predetermined threshold value.

8. The method of claim 7, wherein the threshold value is a function of frame noise of the delta frame.

9. A system for compensating non-uniformities among detector elements of a detector array that captures successive image frames of a scene, where each detector element of the array has at least one neighbor, the system comprising:

(i) an image frame memory that provides a previous image frame, wherein both the previous image frame and a current image frame provided by the detector array are image frames captured by the detector array at different times and are represented as respective matrices each containing image pixel values representing output signals from the detector array;

(ii) a correction unit that applies a plurality of predetermined correction terms to the current frame and the previous frame to generate a corrected current frame and a corrected previous frame;

(iii) a delta-frame unit that generates a delta frame using the current frame and the previous frame;

(iv) a gradient-calculation unit that computes a previous-frame gradient matrix of the corrected previous frame in a first direction, a current-frame gradient matrix of the corrected current frame in a second direction, and a first delta-frame gradient matrix in the first direction and a second delta-frame gradient matrix in the second direction of the delta frame, wherein the second direction is opposite to the first direction;

(v) a gradient-processing unit that computes a first mask using the previous-frame gradient matrix and the first delta-frame gradient matrix, and a second mask using the current-frame gradient matrix and the second delta-frame gradient matrix, wherein the first mask preserves scene-gradient information of a scene associated with the corrected previous frame in the first direction, and the second mask preserves the scene-gradient information associated with the corrected current frame in the second direction;

(vi) a delta-correction unit that computes a plurality of delta-correction values as a function of the previous-frame gradient matrix, the current-frame gradient matrix, and the first and second masks, wherein each of the plurality of delta-correction values is associated with a different one of the detector elements of the array; and

(vii) an update unit that updates each of the plurality of predetermined correction terms with a corresponding delta-correction value.

10. The system of claim 9, wherein the detector array comprises a one-dimensional linear array.

11. The system of claim 9, wherein local scene data is used to compensate for non-uniformities for only a portion of detector elements of the detector array.

12. The system of claim 9, wherein the detector array comprises a two-dimensional planar array.

13. The system of claim 12, wherein the system units (iv) through (vii) perform their respective functions for third and fourth directions perpendicular to the first and second directions, respectively.

14. The system of claim 9, wherein the gradient-processing unit computes the first mask by comparing values of the previous-frame gradient matrix and values of the first delta-frame gradient matrix with a predetermined threshold value.

15. The system of claim 14, wherein the gradient-processing unit computes the second mask by comparing values of the current-frame gradient matrix and values of the second delta-frame gradient matrix with the predetermined threshold value.

16. The system of claim 15, wherein the threshold value is a function of frame noise of the delta frame.