Forming high-dynamic-range (HDR) images using single-channel data

Patent metadata

Field	Value
Publication number	US-9538093-B2
Application number	US-201514663891-A
Country	US
Kind code	B2
Filing date	Mar 20, 2015
Priority date	Sep 20, 2012
Publication date	Jan 3, 2017
Grant date	Jan 3, 2017

Abstract

An image processing method and apparatus are provided. The image processing method includes collecting at least two exposure frames with different brightness in a same scene during different exposure time; combining, for each exposure frame, a raw data unit arranged repeatedly in the exposure frame to obtain first brightness data after the combining; acquiring a correction parameter of all exposure frames according to all first brightness data and performing weighting processing on all the exposure frames by using the correction parameter to obtain a high dynamic range (HDR) image of corrected raw data. The foregoing method can resolve a problem in the prior art that colors, brightness, and contrast of an image obtained from Raw data are severely distorted.

First claim

What is claimed is:

1. An image processing method comprising:
collecting at least two exposure frames with different brightnesses in a same scene during different exposure times;
combining, for each exposure frame, a raw data unit arranged repeatedly in the exposure frame to obtain first brightness data after the combining, wherein the raw data unit comprises raw data that are single-channel data comprising pixels with one of a red (R) value, a green (G) value, and a blue (B) value;
acquiring a correction parameter of all exposure frames according to all first brightness data by:
acquiring a weighting coefficient of each exposure frame according to all the first brightness data;
performing brightness processing on each exposure frame to obtain second brightness data after the brightness processing;
acquiring, according to the weighting coefficient, the second brightness data, and the first brightness data, a correction factor corresponding to all the exposure frames; and
obtaining a correction parameter that comprises the weighting coefficient and the correction factor; and

performing weighting processing on all the exposure frames by using the correction parameter to obtain a high dynamic range (HDR) image of corrected raw data.

2. The method according to claim 1 , wherein performing the weighting processing on all the exposure frames by using the correction parameter to obtain the HDR image of the corrected raw data comprises:
acquiring a weighting coefficient, of each pixel in each exposure frame, corresponding to the exposure frame;
acquiring a product obtained by multiplying the correction factor by each pixel in each exposure frame;
summating products corresponding to the pixels at corresponding locations in all the exposure frames; and
using an image formed by all the pixels obtained after the summating as the HDR image of the corrected raw data.

3. The method according to claim 1 , wherein combining the raw data unit arranged repeatedly in the exposure frame to obtain the first brightness data after the combining comprises:
acquiring a reference brightness value of the exposure frame; and
correcting the reference brightness value by using a preset gamma coefficient to obtain the first brightness data.

4. The method according to claim 3 , wherein acquiring the reference brightness value of the exposure frame comprises:
acquiring an average value of the pixels in the raw data unit; and
using the average value as the reference brightness value.

5. The method according to claim 3 , wherein acquiring the reference brightness value of the exposure frame comprises:
acquiring an average value of green pixels in the raw data unit; and
using the average value as the reference brightness value.

6. The method according to claim 1 , wherein two exposure frames with different brightnesses, which are a first exposure frame R 1 and a second exposure frame R 2 , are collected in a same scene during different exposure times, and wherein acquiring the weighting coefficient of each exposure frame according to all the first brightness data comprises calculating:

w
1
′

=

exp
(

-

I
1
′2

δ
2

)

⁢

⁢

w
2
′

=

exp
(

-

(

255
-

I
2
′

)

2

δ
2

)

w
1

=

w
1
′

w
1
′

+

w
2
′

⁢

⁢

w
2

=

w
2
′

w
1
′

+

w
2
′

;

wherein w′ 1 and w′ 2 represent intermediate values for calculating the weighting coefficient, w 1 represents a weighting coefficient of the first exposure frame, w 2 represents a weighting coefficient of the second exposure frame, δ represents a Gaussian distribution coefficient, I′ 1 represents first brightness data of the first exposure frame, and I′ 2 represents first brightness data of the second exposure frame.

7. The method according to claim 6 , wherein acquiring, according to the weighting coefficient, the second brightness data, and the first brightness data, the correction factor corresponding to all the exposure frames comprises calculating:
α=( I″ 1 *w 1 *I″ 2 *w 2 )/( I′ 1 *w 1 +I′ 2 *w 2 ),

wherein α represents the correction factor, I″ 1 represents second brightness data of the first exposure frame and I″ 2 represents second brightness data of the second exposure frame.

8. The method according to claim 6 , wherein performing the weighting processing on all the exposure frames by using the correction parameter to obtain the HDR image of the corrected raw data comprises calculating R=(R 1 *w 1 +R 2 *w 2 )*α to obtain an HDR image of the corrected raw data when the correction parameter is weighting coefficients w 1 and w 2 and the correction factor, wherein R 1

Assignees

• Huawei Tech Co Ltd

Inventors

• Chen Hai
• Zhang Yan

Classifications

• G06T2207/10016

  Video; Image sequence · CPC title

• G06T2207/10024

  Color image · CPC title

• G06T2207/10144

  Varying exposure · CPC title

• H04N25/60

  Noise processing, e.g. detecting, correcting, reducing or removing noise · CPC title

• H04N23/743

  Bracketing, i.e. taking a series of images with varying exposure conditions · CPC title