Method for generating high dynamic range image from low dynamic range image

1. A method for generating a high dynamic range image from a low dynamic range image, wherein, comprising the following steps:

performing grey-level adjustment on a low dynamic range image to be processed in accordance with a preset mapping relationship to obtain an image after the grey-level adjustment;

selecting a plurality of saturation areas in the image after the grey-level adjustment;

performing grey-level enhancement of the saturation areas in the image after the grey-level adjustment, to obtain a target high dynamic range image; and

outputting the target high dynamic range image; wherein,

the grey-level adjustment comprises inverse-gamma correction and grey-level value increase;

the grey-level value of pixels m the saturation areas is greater than a preset high dynamic saturation threshold; and

the grey-level enhancement comprises the following steps:

generating a smooth enhancement mask;

generating a grey-level enhancement mask in accordance with the smooth enhancement mask; and

enhancing the grey-level value of the saturation areas in the image after the grey-level adjustment from a range of (T HDR , Value 1 ) to a range of (T HDR , Value) using the grey-level enhancement mask, so as to obtain a target high dynamic range image,

wherein, T HDR is a high dynamic range saturation threshold, Value 1 is a maximum grey-level value in the image after the grey-level adjustment, Value is a maximum grey-level value in the target high dynamic range image.

2. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 1 , wherein, performing grey-level adjustment in accordance with the following formula:

C

⁡

(

i

)

=

(

2

n

-

1

)

⁡

[

i

2

m

-

1

]

1

γ

wherein, C(i) is a grey-level value of each pixel in the image after the grey-level adjustment, i is a grey-level value of each pixel in the low dynamic range image to be processed, and i∈[0,255];

m is the number of grey level of the low dynamic range image to be processed, n IS the number of grey level of the image after the grey-level adjustment, and m≤n; and

γ is a preset gamma value.

3. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 1 , wherein, performing grey-level adjustment in accordance with the following formula:

C

⁡

(

i

)

=

{

(

(

2

n

-

1

)

⁡

[

i

2

m

-

1

]

1

γ

,

i

≥

γ

8

⁢

2

n

k

·

i

,

i

<

γ

8

⁢

2

n

;

⁢

k

=

C

⁡

(

γ

8

⁢

2

n

)

γ

8

⁢

2

n

;

wherein, C(i) is a grey-level value of each pixel in the image after the grey-level

adjustment, i is a grey-level value of each pixel in the low dynamic range image to be processed, and i∈[0,255];

m is the number of grey level of the low dynamic range image to be processed, n is the number of grey level of the image after the grey-level adjustment, and m≤n; and

γ is a preset gamma value.

4. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 1 , wherein, performing grey-level adjustment in accordance with the following formula:

C

⁡

(

i

)

=

{

(

2

n

-

1

)

⁡

[

i

2

m

-

1

]

1

γ

,

i

≥

γ

8

⁢

2

n

(

k

·

i

+

(

2

n

-

1

)

⁡

[

i

2

m

-

1

]

1

r

)

/

2

,

i

<

γ

8

⁢

2

n

;

⁢

k

=

C

⁡

(

γ

8

⁢

2

n

)

γ

8

⁢

2

n

;

wherein, C(i) is a grey-level value of each pixel in the image after the grey-level adjustment, i is a grey-level value of each pixel in the low dynamic range image to be processed, and i∈[0,255];

m is the number of grey level of the low dynamic range image to be processed, n is the number of grey level of the image after the grey-level adjustment, and m≤n; and

γ is a preset gamma value.

5. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 1 , wherein the grey-level adjustment further comprises grey-level compression, and performing grey-level adjustment in accordance with the following formula:

C

⁡

(

i

)

=

{

β

·

(

2

n

-

1

)

⁡

[

i

2

m

-

1

]

1

γ

,

i

≥

γ

8

⁢

2

n

β

·

k

·

i

,

i

<

γ

8

⁢

2

n

;

⁢

k

=

C

⁡

(

γ

8

⁢

2

n

)

γ

8

⁢

2

n

;

wherein, C(i) is a grey-level value of each pixel in the image after the grey-level adjustment, i is a grey-level value of each pixel in the low dynamic range image to be processed, and i∈[0,255];

m is the number of grey level of the low dynamic range image to be processed, n is the number of grey level of the image after the grey-level adjustment, and m≤n; and

γ is a preset gamma value, β is a preset compression factor, and 0<β≤1.

6. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 1 , wherein the grey-level adjustment further comprises grey-level compression, and performing grey-level adjustment in accordance with the following formula:

C

⁡

(

i

)

=

{

β

·

(

2

n

-

1

)

⁡

[

i

2

m

-

1

]

1

γ

,

i

≥

γ

8

⁢

2

n

β

·

(

k

·

i

+

(

2

n

-

1

)

⁡

[

i

2

m

-

1

]

1

r

)

/

2

,

i

<

γ

8

⁢

2

n

;

⁢

k

=

C

⁡

(

γ

8

⁢

2

n

)

γ

8

⁢

2

n

;

wherein, C(i) is a grey-level value of each pixel in the image after the grey-level adjustment, i is a grey-level value of each pixel in the low dynamic range image to be processed, and i∈[0,255];

m is the number of grey level of the low dynamic range image to be processed, n is the number of grey level of the image after the grey-level adjustment, and m≤n; and

γ is a preset gamma value, β is a preset compression factor, and 0<β≤1.

7. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 6 , wherein the grey-level enhancement comprises the following steps:

generating a smooth enhancement mask;

generating a grey-level enhancement mask in accordance with the smooth enhancement mask;

enhancing the grey-level value of the saturation areas in the image after the grey-level adjustment from a range of (T HDR , Value 1 ) to a range of (T HDR , Value) using the grey-level enhancement mask, so as to obtain a target high dynamic range image;

wherein, T HDR is a high dynamic range saturation threshold, Value 1 is a maximum grey-level value in the image after the grey-level adjustment, Value is a maximum grey-level value in the target high dynamic range image.

8. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 1 , wherein the generating a smooth enhancement mask comprises the following steps:

performing binarized segmentation of the image after the grey-level adjustment with the preset high dynamic saturation threshold, so as to obtain a corresponding binarized image;

generating a Gaussian pyramid in accordance with the binarized image, and defining a smooth enhancement mask as being equal to the Gaussian pyramid; and

performing multiple Gaussian blur to the smooth enhancement mask, and outputting the smooth enhancement mask.

9. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 8 , wherein performing multiple Gaussian blur to the smooth enhancement mask comprises the following steps:

assuming i=k, performing Gaussian blur to the smooth enhancement mask;

deducting 1 from the value of i after each Gaussian blur;

conducting upsampling to the smooth enhancement mask, and performing Gaussian blur repeatedly to the smooth enhancement mask, if i>0; and

outputting the smooth enhancement mask, if i≤0.

10. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 8 , wherein the grey-level enhancement further comprises the step of generating an edge enhancement mask, and generating the grey-level enhancement mask in accordance with the smooth enhancement mask and the edge enhancement mask.

11. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 10 , wherein the generating an edge enhancement mask comprises the following steps:

selecting a seed pixel in each of the saturation areas in accordance with the Gaussian pyramid;

performing the flood fill algorithm in the image with the seed pixel after the smooth enhancement, and calculating an edge stopping mask;

performing pyramid subsampling to the edge stopping mask; and

performing multiple dilation and opening operations to the edge enhancement mask, with the edge enhancement mask to be equal to the Gaussian pyramid, and outputting the edge enhancement mask.

12. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 11 , wherein performing multiple dilation and opening operations to the edge enhancement mask comprises the following steps:

assuming i=k, performing dilation to the edge enhancement mask;

updating the edge enhancement mask in accordance with the following formula:

s _mask= e _mask & s _mask;

wherein, s_mask is the edge enhancement mask, e_mask is the edge stopping mask;

performing an opening operation to the edge enhancement mask;

deducting 1 from the value of i after each opening operation;

performing upsampling of the edge enhancement mask, and repeating dilation, if i>0; and

outputting the edge enhancement mask, if i≤0.

13. The method for generating a high dynamic range image from a low dynamic range image in accordance with claim 11 , wherein the generating a grey-level enhancement mask comprises the following steps:

generating the grey-level enhancement mask m accordance with the following formula:

mask= b _mask* s _mask;

wherein, mask is the grey-level enhancement mask, b_mask is the smooth enhancement mask, and s_mask is the edge enhancement mask;

normalizing the grey-level enhancement mask as [1, α], wherein α=1/β, and β is a preset compression factor; and

outputting the grey-level enhancement mask.