Bias and plateau limited advanced contrast enhancement

• US 8,774,554 B1
• Filed: 01/18/2012
• Issued: 07/08/2014
• Est. Priority Date: 05/09/2011
• Status: Active Grant

• Current Assignee: Elbit Systems of America LLC (Elbit Systems Limited)
• Original Assignee: Exelis Incorporated (L3Harris Technologies, Inc.)
• Inventors: Winn, Stuart Martin
• Primary Examiner(s): Liu, Li
• Application Number:
  US13/352,765
• Time in Patent Office: 902 Days
• Field of Search: None
• US Class Current: 382/274
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20012 Locally adaptive
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 5/40 using histogram techniques
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 5/94 based on local image properties, e.g. for local contrast enhancement

Abstract

A method of enhancing an image. This includes the steps of: dividing the image into multiple tiles and constructing a histogram for each tile, wherein the histogram represents a distribution of pixel intensity values in each tile. In addition, the method applies a bias value to the histogram and a plateau value to the histogram. A cumulative histogram is constructed after applying the bias value and the plateau value to the histogram. The method transforms each pixel in the image, by weighing four cumulative histograms, respectively, in four adjacent and closest tiles to the pixel under consideration. The pixel under consideration is modified based on a weighted cumulative histogram.

Claims

1. A method of enhancing an image comprising the steps of:

dividing a previous image into multiple tiles;

constructing a histogram for each tile, wherein the histogram represents a distribution of pixel intensity values in each tile;

applying a bias value to the histogram;

applying a plateau value to the histogram;

constructing a cumulative distribution function (CDF) for each tile, after applying the bias value and the plateau value to the histogram;

receiving a present image;

determining an intensity value for a pixel in the present image;

finding four adjacent tiles in the previous image that surround the pixel in the present image;

transforming the intensity value of the pixel in the present image based on respective CDFs constructed for the four adjacent tiles in the previous image to obtain four intensity levels; and

using bilinear interpolation to calculate a single intensity value of the pixel in the present image based on the four intensity levels.

2. The method of claim 1 wherein dividing the previous image into multiple tiles includes:

dividing the previous image into a plurality of tiles, wherein each tile is a region in the previous image of n×

m pixels, where n and m are integer values.

3. The method of claim 1 wherein applying the bias value includes:

determining a gain value in the previous image; and

setting the bias value, based on the gain value of the previous image.

4. The method of claim 3 wherein

the bias value is increased, if the gain value exceeds a predetermined value, and

the bias value is decreased, if the gain value is below the predetermined value.

5. The method of claim 1 wherein applying the plateau value includes:

clipping the histogram at a predetermined level, in which the clipping spans across the distribution of pixel intensity values per tile.

6. The method of claim 5 wherein the predetermined level is dependent upon type of sensor, including a night vision goggle (NVG) sensor.

7. The method of claim 1 wherein using bilinear interpolation includes:

finding distances, in an X-direction, from the pixel in the current image to the centers of the four tiles to form two intensity values, and

finding a distance, in a Y-direction, from the pixel in the current image to the two intensity values to form one intensity value.

8. The method of claim 1 wherein the step of constructing the histogram includes the following:

dividing the histogram into two portions, wherein

a first portion includes a count of pixels spanning between 0 and 50 percent of maximum pixel intensity, and

a second portion includes a count of pixels spanning between 50 and 100 percent of maximum pixel intensity.

9. A system comprising:

an image intensifier for amplifying a low level light from a scene,

a processor for receiving the amplified low level light and forming an enhanced image of the scene, and

a display for viewing the enhanced image;

wherein the processor executes the following steps;

forming an image from the amplified low level light,

dividing a previous image into multiple tiles;

constructing a histogram for each tile, wherein the histogram represents a distribution of pixel intensity values in each tile;

applying a bias value to the histogram;

applying a plateau value to the histogram;

constructing a cumulative distribution function (CDF) for each tile, after applying the bias value and the plateau value to the histogram;

receiving a present image;

determining an intensity value for a pixel in the present image;

finding four adjacent tiles in the previous image that surround the pixel in the present image;

transforming the intensity value of the pixel in the present image based on respective CDFs constructed for the four adjacent tiles in the previous image to obtain four intensity levels; and

using bilinear interpolation to calculate a single intensity value of the pixel in the present image based on the four intensity levels.

10. The system of claim 9 wherein the processor includes a module for applying the bias value by executing the following steps:

determining a gain value in the image; and

setting the bias value based on the gain value in the image; and

wherein

the bias value is increased, if the gain value exceeds a predetermined value, and

the bias value is decreased if the gain value is below the predetermined value.

11. The system of claim 9 wherein the processor includes a module for applying the plateau value by executing the following step:

clipping the histogram at a predetermined level, in which the clipping spans across the distribution of pixel intensity values per tile;

wherein the predetermined level is dependent upon a type of sensor, including an NVG sensor.