Image processing method, electronic device, and readable storage medium

1 . An image processing method, comprising:

obtaining an original image;

denoising the original image using an image filtering algorithm to form a preprocessed image S;

performing contrast enhancement in the preprocessed image S to form a contrast-enhanced image J;

adjusting saturation of the contrast-enhanced image J to form an enhanced display image M;

sharpening the enhanced display image M to form a sharpened image N; and

performing dynamic range enhancement on the sharpened image N using a dynamic range image enhancement algorithm to form an output image,

wherein the step “sharpening the enhanced display image M to form a sharpened image N” comprises:

step S 51 , extracting three-channel images light c of a brightness layer image for R, G, B three-channel images of the enhanced display image M using a guided filtering algorithm;

step S 52 , obtaining R, G, B three-channel images N c corresponding to the sharpened image N based on the three-channel images light c of the brightness layer image;

N

c

(

i

,

j

)

=

M

c

(

i

,

j

)

+

(

M

c

(

i

,

j

)

-

light

c

(

i

,

j

)

)

*

α

,

wherein, c=[R,G,B], representing one of three channels R, G, B, α is a constant, M c (i,j) represents the value of the pixel at coordinates (i,j) in one of the R, G, B three-channel images corresponding to the enhanced display image M, N c (i,j) represents the value of the pixel at coordinates (i,j) in one of the R, G, B three-channel images corresponding to the sharpened image N, and light c (i,j) represents the value of the pixel at coordinates (i,j) in one of the three-channel images of the brightness layer image;

step S 53 , merging three channels of the R, G, B three-channel images N c to form the sharpened image N;

wherein the method further comprises:

downsampling the enhanced display image M before step S 51 to form a downsampled image;

replacing the enhanced display image M with the downsampled image when performing step S 51 ; and

after step S 51 , upsampling the three-channel images light c formed in step S 51 to form three-channel images light c with the same size as the original image.

2 . The image processing method of claim 1 , wherein the step “denoising the original image using an image filtering algorithm to form a preprocessed image S” comprises:

taking a Gaussian filtering algorithm or a mean filtering algorithm as the image filter algorithm, representing the value of each pixel in the preprocessed image S as S (i,j),

S ( i,j )=blur( I ( i,j ),size( r,r )),

wherein, blur represents filtering function, I(i,j) represents the value of the pixel at coordinates (i,j) in the original image, size (r,r) represents the size of the filtering algorithm, r represents a filtering algorithm parameter, r is a constant, and its numerical value is related to the size of the original image and the noise level of the original image.

3 . The image processing method of claim 1 , wherein the step “performing contrast enhancement in the preprocessed image S to form a contrast-enhanced image J” comprises: performing a gamma transformation on the preprocessed image S to form a gamma image J 1 ; converting the preprocessed image S to a grayscale image S 1 , calculating the mean value of grayscale values of the grayscale image S 1 to form a brightness value m 1 corresponding to the preprocessed image S;

converting the gamma image J 1 to a grayscale image J 2 , obtaining pixels in the grayscale image J 2 with grayscale values greater than a preset first grayscale threshold, calculating the mean value of grayscale values of the obtained pixels to form a brightness value m 2 corresponding to the gamma image J 1 ;

obtaining the contrast-enhanced image J based on the gamma image J 1 , the brightness value m 1 corresponding to the preprocessed image S, and the brightness value m 2 corresponding to the gamma image J 1 ;

J

⁡

(

i

,

j

)

=

β

*

J

⁢

1

⁢

(

i

,

j

)

,

β

=

{

1

,

m

⁢

2

≥

m

⁢

1

min

⁢

(

m

⁢

1

,

m

⁢

2

+

v

)

m

⁢

2

,

m

⁢

2

<

m

⁢

1

,

wherein v is a constant, J 1 ( i,j ) represents the value of the pixel at coordinates (i,j) in the gamma image J 1 , and J (i,j) represents the value of the pixel at coordinates (i,j) in the contrast-enhanced image J.

4 . The image processing method of claim 1 , wherein the step “adjusting saturation of the contrast-enhanced image J to form an enhanced display image M” comprises:

obtaining R, G, B three-channel images corresponding to the contrast-enhanced image J, and for each pixel in the contrast-enhanced image J, taking the maximum value of the pixel in the R, G, B three-channel images as the grayscale value to form a maximum value image M max , and taking the minimum value of the pixel in the R, G, B three-channel images as the grayscale value to form a minimum value image M min ;

obtaining R, G, B three-channel images M c corresponding to the enhanced display image M based on the maximum value image M max and the minimum value image M min ;

M

c

⁢

(

i

,

j

)

=

L

+

(

J

c

(

i

,

j

)

-

L

)

*

(

1

+

K

)

,

L

=

M

⁡

(

i

,

j

)

max

+

M

⁡

(

i

,

j

)

min

2

;

wherein, c=[R,G,B], representing one of three channels R, G, B, K is a constant, M c (i,j) represents the value of the pixel at coordinates (i, j) in one of the R, G, B three-channel images corresponding to the enhanced display image M, J c (i, j) represents the value of the pixel at coordinates (i, j) in one of the R, G, B three-channel images corresponding to the contrast-enhanced image J, M (i,j) max represents the value of the pixel at coordinates (i, j) in the maximum value image M max , and M(i, j) min represents the value of the pixel at coordinates (i, j) in the minimum value image M min ;

merging three channels of the R, G, B three-channel images M c to form the enhanced display image M.

5 . The image processing method of claim 1 , wherein the step “sharpening the enhanced display image M to form a sharpened image N” comprises:

sharpening the enhanced display image M using a Laplacian pyramid algorithm to form the sharpened image N.

6 . The image processing method of claim 1 , wherein the dynamic range image enhancement algorithm is local histogram enhancement algorithm or Retinex enhancement algorithm.

7 . The image processing method of claim 1 , wherein the step “performing dynamic range enhancement on the sharpened image N using a dynamic range image enhancement algorithm to form an output image” comprises:

step S 61 , obtaining R, G, B three-channel images of the sharpened image N, and for each pixel in the sharpened image N, taking the maximum value corresponding in the R, G, B three-channel images as the grayscale value to form an initial image L 1 ;

step S 62 , performing mean filtering on the initial image L 1 to form a light image L;

step S 63 , calculating the total number of pixels P 1 with grayscale values less than a preset second grayscale threshold in the light image L, wherein

if P 1 ≤P, the sharpened image N is used as the output image E; and

if P 1 >P, performing steps S 64 , S 65 , and S 66 sequentially; wherein P is a preset determination threshold, which is a constant;

step S 64 , calculating a mean grayscale value mean of the pixels with grayscale values less than the preset second grayscale threshold in the light image L; and

obtaining an exposure image G according to the mean grayscale value mean,

G

⁡

(

i

,

j

)

=

beta

*

N

⁡

(

i

,

j

)

q

,

beta

=

exp

⁢

(

(

1

-

q

)

*

b

)

,

q

=

k

a

;

k

=

135

/

mean

,

k

=

max

⁢

(

1

,

min

⁢

(

k

,

7

)

)

;

wherein, a and b are constants, G (i,j) represents the value of the pixel at coordinates (i,j) in the exposure image G, and N (i,j) represents the value of the pixel at coordinates (i,j) in the sharpened image N;

step S 65 , obtaining R, G, B three-channel images E c corresponding to the output image E based on the exposure image G and the sharpened image N;

E

c

(

i

,

j

)

=

N

c

(

i

,

j

)

*

ω

+

G

c

(

i

,

j

)

*

(

1

-

ω

)

,

ω

=

1

/

L

⁡

(

i

,

j

)

f

;

f

=

P

⁢

1

w

*

h

;

wherein, c=[R,G,B], representing one of three channels R, G, and B, P 1 is a constant, E c (i,j) represents the value of the pixel at coordinates (i, j) in one of the R, G, B three-channel images corresponding to the output image E, N c (i,j) represents the value of the pixel at coordinates (i, j) in one of the R, G, B three-channel images corresponding to the sharpened image N, G c (i,j) represents the value of the pixel at coordinates (i, j) in one of R, G, B three-channel images corresponding to the exposure image G, L (i,j) represents the value of the pixel at coordinate (i,j) in the light image L; w and h represent the width and height of the sharpened image N, respectively; f is the proportion of the total number of pixels P 1 with grayscale values less than the preset second grayscale threshold in the light image L; ω is a weight coefficient, which is normalized to a value between 0 and 1;

step S 66 , merging three channels of the R, G, B three-channel images E c to form the output image E.

8 . An electronic device, comprising a memory and a processor, wherein the memory stores a computer program that runs on the processor, and the processor executes the program to implement steps of the image processing method, wherein the image processing method comprises:

obtaining an original image;

denoising the original image using an image filtering algorithm to form a preprocessed image S;

performing contrast enhancement in the preprocessed image S to form a contrast-enhanced image J;

adjusting saturation of the contrast-enhanced image J to form an enhanced display image M;

sharpening the enhanced display image M to form a sharpened image N; and

performing dynamic range enhancement on the sharpened image N using a dynamic range image enhancement algorithm to form an output image,

wherein the step “sharpening the enhanced display image M to form a sharpened image N” comprises:

step S 51 , extracting three-channel images light c of a brightness layer image for R, G, B three-channel images of the enhanced display image M using a guided filtering algorithm;

step S 52 , obtaining R, G, B three-channel images N c corresponding to the sharpened image N based on the three-channel images light c of the brightness layer image;

N

c

(

i

,

j

)

=

M

c

(

i

,

j

)

+

(

M

c

(

i

,

j

)

-

light

c

(

i

,

j

)

)

*

α

,

wherein, c=[R,G,B], representing one of three channels R, G, B, α is a constant, M c (i,j) represents the value of the pixel at coordinates (i,j) in one of the R, G, B three-channel images corresponding to the enhanced display image M, N c (i,j) represents the value of the pixel at coordinates (i,j) in one of the R, G, B three-channel images corresponding to the sharpened image N, and light c (i,j) represents the value of the pixel at coordinates (i,j) in one of the three-channel images of the brightness layer image;

step S 53 , merging three channels of the R, G, B three-channel images N c to form the sharpened image N;

wherein the method further comprises:

downsampling the enhanced display image M before step S 51 to form a downsampled image;

replacing the enhanced display image M with the downsampled image when performing step S 51 ; and

after step S 51 , upsampling the three-channel images light c formed in step S 51 to form three-channel images light c with the same size as the original image.