High dynamic range image synthesis and tone mapping method and electronic device

Abstract

The present application relates to the technical field of image processing, in particular to a high dynamic range image synthesis and tone mapping method and electronic equipment, the present application obtains multiple RAW images with different exposure under the same scene through a camera; the RAW images with different exposure are fused through a square weighting function to obtain a high dynamic range image synthesized in the RAW domain; the high dynamic range image is subjected to channel tone mapping to obtain a low dynamic range image in the RAW domain. The present application solves the problem of synthesizing and mapping a high dynamic range image on RGB, which occupies a large computer resource and seriously affects the color and brightness information of the image. The present application has fast calculation speed, low resource consumption, good result perception and is suitable for hardware integrated high dynamic range image synthesis, and has strong market application prospect.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to a high dynamic range image synthesis and tone mapping method and electronic device. Background Technology

[0002] The dynamic range of an image refers to the ratio between its maximum and minimum grayscale values. High dynamic range (HDR) images provide more detail than ordinary images. Synthesizing HDR images from low dynamic range (LVR) images with different exposure times can better reproduce the realistic scene perceived by the human eye. Since the dynamic range of HDR images is far greater than that of typical displays, a tone mapping technique is needed to map HDR images to low dynamic range (LVR) images.

[0003] Currently, most cameras use CMOS or CCD image sensors to capture light source signals, which are then converted into raw digital data, i.e., RAW data. After image signal processing, images with good visual effects are generated.

[0004] Most existing algorithms synthesize and map high dynamic range images onto RGB, which consumes significant computer resources. Simply applying these methods to RAW data processing can severely impact image color and brightness information. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention discloses a high dynamic range image synthesis and tone mapping method and electronic device to solve the aforementioned problems.

[0006] This invention is achieved through the following technical solution:

[0007] In a first aspect, the present invention provides a high dynamic range image synthesis and tone mapping method, comprising the following steps:

[0008] The S1 captures multiple RAW images with different exposures of the same scene using a camera.

[0009] S2 uses a squared weighting function to perform exposure fusion on RAW images with different exposures, resulting in a high dynamic range image synthesized in the RAW domain;

[0010] S3 performs channel-by-channel tone mapping on the high dynamic range image to obtain a low dynamic range image in the RAW domain.

[0011] Furthermore, in the method, the RAW image is the raw data of the light source signal captured by the camera's CMOS or CCD image sensor and converted into a digital signal.

[0012] Furthermore, in the method, well-exposed areas are assigned higher weights during fusion, and then weighted synthesis is performed, such that the irradiance value and pixel value of the image satisfy the following formula:

[0013] Z i j = f(E i *t j )

[0014] Where the subscript i represents the i-th pixel, the subscript j represents the j-th image, and t j Let E represent the exposure time of the j-th image, zij represent the actual pixel value of the i-th pixel in the j-th image, and Eij represent the actual pixel value of the i-th pixel in the j-th image. i This represents the irradiance value of the i-th pixel.

[0015] Furthermore, in the method described above, when synthesizing RAW format images, the camera response function f is expressed as: f = c

[0016] Where c represents a constant, and E is the irradiance value of the image. ij This can be expressed as:

[0017] ,

[0018] t j Z represents the exposure time of the j-th image. ij This represents the pixel value of the i-th pixel in the j-th image;

[0019] After obtaining the irradiance values ​​of the image, image synthesis is performed. A quadratic function with a rapid gradient increase or decrease is used at locations with smaller or larger pixels. Specifically:

[0020] ,

[0021] Where bit represents the number of bits in the input image, Z ij Let w(Z) represent the pixel value of the i-th pixel in the j-th image. ij ) represents the weight of the i-th pixel in the j-th image.

[0022] Furthermore, in the method,

[0023] Assume the exposure time ratio Δt is:

[0024] ,

[0025] Among them, t j t1 represents the exposure time of the j-th image, and t1 represents the exposure time of the 1-th image;

[0026] Let the image brightness L be the maximum pixel value within an RGGB cell. If the brightness values ​​within the same cell are the same, a pixel is considered overexposed if the following condition is met:

[0027] ,

[0028] Where black represents the black level value, bit represents the number of bits in the input image, m is an adjustable value, Δt represents the exposure time ratio, and L... i1 L represents the brightness of the i-th pixel in the first image. ij This represents the brightness of the i-th pixel in the j-th image. Based on the above conditions, overexposed areas are identified, and their weights are set to 0, meaning they are not included in the fusion process. The final fusion result is as follows:

[0029] ,

[0030] Where C is related to the final synthesized bit depth, assuming outbit represents the desired bit depth of the synthesized image, then C is specifically expressed as:

[0031] C = 2(outbit - bit).

[0032] Furthermore, in the method, the high dynamic range image is subjected to channel-specific tone mapping to obtain a low dynamic range image in the RAW domain. This involves extracting the RGB data from the RAW image separately, mapping them using the algorithms described below, and finally combining the three channels of data to obtain the final output image.

[0033] Furthermore, the method involves performing channel-specific global processing on the high dynamic range image, including extracting the RGB data from the RAW image separately.

[0034] ,

[0035] ,

[0036] ,

[0037] Where c represents the color adaptation parameter, a represents the brightness adaptation parameter, L represents the grayscale image of the input RAW image obtained by linear interpolation, and Lav represents the mean of the grayscale image, where I r|g|b and These represent the RGB values ​​and the pixel mean of the three channels of the RAW image after white balance.

[0038] This represents the pixel value after weighting the pixel brightness point by point.

[0039] I represents the pixel value after weighting by the average pixel brightness. a_r|g|b This represents the pixel value after global processing.

[0040] Furthermore, in the method, during the image compression process, it is necessary to adjust the overall brightness and contrast of the image. The brightness adjustment parameter f is obtained exponentially, specifically expressed as follows:

[0041] f = exp(f')

[0042] f' represents a brightness adjustment parameter that can be manually entered. The larger the f', the higher the brightness, but the image contrast will also be relatively reduced.

[0043] The contrast parameter m is obtained using an adaptive method, and the specific calculation formula is as follows:

[0044] m = 0.3 + 0.7 * k 1.4 ,

[0045] ,

[0046] k represents the image key value, L max L represents the maximum value of the grayscale image obtained through interpolation. min L represents the minimum value. av The mean of the entire grayscale image.

[0047] Furthermore, in the method, the high dynamic range image is subjected to channel-by-channel tone mapping to obtain a low dynamic range image in the RAW domain, including:

[0048] σ(I a_r|g|b )=(f*I a_r|g|b ) m ,

[0049] ,

[0050] I r|g|b V represents the number of pixels in the input image. r|g|b This represents the output data after color level mapping, σ(I) a_r|g|b ) represents the half-saturation constant, I a_r|g|b This represents the pixel value after global processing, where f is the brightness adjustment parameter and m is the contrast parameter. V r|g|b The images are combined to form a RAW image V in RGGB format. Finally, normalization is performed to obtain the globally mapped RAW image. The normalization method is as follows:

[0051] ,

[0052] Where V max V represents the maximum value of the grayscale image V in the output image. min V represents the minimum value. out This indicates the output RAW image.

[0053] In a second aspect, the present invention provides an electronic device, including a processor; and a memory storing computer-readable instructions that, when executed by the processor, implement the high dynamic range image synthesis and tone mapping method according to the first aspect.

[0054] The beneficial effects of this invention are as follows:

[0055] This invention solves the problem that synthesizing and mapping high dynamic range images on RGB consumes a large amount of computer resources and seriously affects the color and brightness information of the image. This invention has fast calculation speed, low resource consumption, good visual results, and is suitable for hardware-integrated high dynamic range image synthesis, and has a strong market application prospect. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a step diagram of a high dynamic range image synthesis and tone mapping method according to an embodiment of the present invention;

[0058] Figure 2 This is a flowchart illustrating the steps of a high dynamic range image synthesis and tone mapping algorithm disclosed in an embodiment of the present invention;

[0059] Figure 3 This is a diagram of the RAW data format disclosed in an embodiment of the present invention. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Example 1

[0062] This embodiment provides a high dynamic range image synthesis and tone mapping method, such as Figure 1 As shown, it includes the following steps:

[0063] The S1 captures multiple RAW images with different exposures of the same scene using a camera.

[0064] S2 uses a squared weighting function to perform exposure fusion on RAW images with different exposures, resulting in a high dynamic range image synthesized in the RAW domain;

[0065] S3 performs channel-by-channel tone mapping on the high dynamic range image to obtain a low dynamic range image in the RAW domain.

[0066] In this embodiment, the RAW image is the raw data of the light source signal captured by the camera's CMOS or CCD image sensor and converted into a digital signal.

[0067] This embodiment is based on the linear characteristics of RAW images, including:

[0068] The pixel size of an image in the RAW domain can directly reflect the irradiance of the real scene. This means that, unlike processing images in the RGB domain, there is no need to predict and fit the camera response curve. The camera response function can be understood as a constant.

[0069] Exposure fusion of RAW images with different exposures is performed using a squared weighting function to obtain a synthesized high dynamic range image in the RAW domain, including:

[0070] During compositing, we aim to assign higher weights to well-exposed areas and lower weights to poorly exposed areas. To ensure that underexposed areas utilize more pixels from the long-exposure image, we employ a quadratic function with a rapid gradient increase (decrease) at smaller (larger) pixel locations. Specifically:

[0071] ,

[0072] Where bit represents the number of bits in the input image, Z ij Let w(Z) represent the pixel value of the i-th pixel in the j-th image. ij Let represent the weight of the i-th pixel in the j-th image. For overexposed locations, the pixel value and exposure time do not have a linear relationship, so overexposed locations are not included in the fusion process.

[0073] In this embodiment, for overexposed locations, the pixel value and exposure time do not satisfy a linear relationship, and the overexposed locations do not participate in the fusion process, including:

[0074] Based on this condition, determine whether it is an overexposed area, assuming the exposure time ratio Δt is:

[0075] ,

[0076] t jt1 represents the exposure time of the j-th image, and t1 represents the exposure time of the 1-th image.

[0077] The image brightness L is the maximum pixel value within an RGGB cell block. Pixels with the same brightness value within the same cell block are considered overexposed when the following conditions are met:

[0078] ,

[0079] Where black represents the black level value, bit represents the number of bits in the input image, m is a very small adjustable value, Δt represents the exposure time ratio, and L... i1 L represents the brightness of the i-th pixel in the first image. ij This represents the brightness of the i-th pixel in the j-th image.

[0080] This embodiment performs channel-specific global processing on high dynamic range images, including:

[0081] Extract the RGB data from the RAW image separately.

[0082] ,

[0083] ,

[0084] ,

[0085] Where c represents the color adaptation parameter, a represents the brightness adaptation parameter, and L represents the grayscale image of the input RAW image, obtained using linear interpolation. av Represents the mean of the grayscale image, where I r|g|b and These represent the RGB values ​​and the pixel mean of the three channels of the RAW image after white balance.

[0086] This represents the pixel value after weighting the pixel brightness point by point.

[0087] I represents the pixel value after weighting by the average pixel brightness. a_r|g|b This represents the pixel value after global processing.

[0088] This embodiment performs channel-wise tone mapping on a high dynamic range image to obtain a low dynamic range image in the RAW domain, including:

[0089] σ(I a_r|g|b )=(f*I a_r|g|b ) m ,

[0090] ,

[0091] I r|g|b V represents the number of pixels in the input image. r|g|b This represents the output data after color level mapping, σ(I) a_r|g|b ) represents the half-saturation constant, I a_r|g|b This represents the pixel value after global processing, where f is the brightness adjustment parameter and m is the contrast parameter. V r|g|b The images are combined to form a RAW image V in RGGB format. Finally, normalization is performed to obtain the globally mapped RAW image. The normalization method is as follows:

[0092] ,

[0093] Where V max V represents the maximum value of the grayscale image V in the output image. min V represents the minimum value. out This indicates the output RAW image.

[0094] This embodiment solves the problem that synthesizing and mapping high dynamic range images on RGB consumes a large amount of computer resources and seriously affects the color and brightness information of the image.

[0095] Example 2

[0096] At the application level, this embodiment provides a high dynamic range image synthesis and tone mapping algorithm, such as... Figure 2 As shown, it includes the following steps:

[0097] Step 1: Acquire multiple RAW images of the same scene with different exposures;

[0098] Step 2: Based on the linear characteristics of RAW images, a squared weighting function is used to perform exposure fusion on RAW images with different exposures to obtain a high dynamic range image synthesized in the RAW domain;

[0099] Step 3: Perform channel-by-channel tone mapping on the high dynamic range image to obtain a low dynamic range image in the RAW domain;

[0100] This embodiment acquires multiple RAW images of the same scene with different exposures, including the following steps: The synthesis of high dynamic range images requires multiple RAW images of the same scene with different exposures, which can be two, three, or more exposures. RAW data is acquired from the camera, that is, the raw data of the light source signals captured by the camera's CMOS or CCD image sensor and converted into digital signals.

[0101] This embodiment, based on the linear characteristics of RAW images, uses a squared weighting function to fuse RAW images with different exposures, obtaining a synthesized high dynamic range image in the RAW domain. The steps include: images with different exposures contain different well-exposed areas; these well-exposed areas are assigned higher weights during fusion, and then weighted synthesis yields an image containing more detail and a higher dynamic range. The irradiance value and pixel value of the image satisfy the following formula:

[0102] Z ij =f(E i *t j ),

[0103] Where the subscript i represents the i-th pixel, the subscript j represents the j-th image, and t j z represents the exposure time of the j-th image. ij E represents the true pixel value of the i-th pixel in the j-th image. i This represents the irradiance value of the i-th pixel.

[0104] Since this embodiment studies the synthesis of RAW format images, RAW images are images directly acquired by the camera without nonlinear processing. RAW images preserve the most original and complete data information, and they also have an important property: linearity. That is, under the premise that the brightness of the scene does not change in the same static scene, the pixel value of the image matrix in the RAW domain is proportional to the camera's exposure time. Of course, this is under the premise that the pixels are not overexposed. This property is very important for subsequent research.

[0105] This property indicates that the pixel values ​​of an image in the RAW domain can directly reflect the irradiance, i.e., brightness, of objects in the real scene. This means that it is not necessary to predict and fit the camera's camera response curve as is done when processing images in the RGB domain.

[0106] Therefore, the camera response function can be understood as a constant, that is, the camera response function f is expressed as:

[0107] f = g, where g represents a constant;

[0108] c represents a constant, which is defaulted to 1.0 in practice. Therefore, the image's irradiance value E... ij This can be expressed as:

[0109] ,

[0110] t j Z represents the exposure time of the j-th image. ij This represents the pixel value of the i-th pixel in the j-th image.

[0111] After obtaining the irradiance value of the image in this embodiment, image compositing can be performed. During compositing, we aim to assign higher weights to well-exposed areas and lower weights to poorly exposed areas. To ensure that underexposed areas utilize more pixels from the long-exposure image, a quadratic function is used where the gradient rises faster at smaller pixel locations and falls faster at larger pixel locations. Specifically:

[0112] ,

[0113] Where bit represents the number of bits in the input image, Z ij Let w(Z) represent the pixel value of the i-th pixel in the j-th image. ij Let represent the weight of the i-th pixel in the j-th image. For overexposed locations, the pixel value and exposure time do not have a linear relationship. If overexposed pixels are still included in the fusion process, the image will exhibit color cast. Therefore, in the experiment, it is desirable that overexposed locations are not included in the fusion process. Assume the exposure time ratio Δt is: ,

[0114] t j t1 represents the exposure time of the j-th image, and t1 represents the exposure time of the 1-th image.

[0115] The image brightness L is the maximum pixel value within an RGGB cell block. Pixels with the same brightness value within the same cell block are considered overexposed when the following conditions are met:

[0116] ,

[0117] Where black represents the black level value, bit represents the number of bits in the input image, m is a very small adjustable value, Δt represents the exposure time ratio, and L... i1 L represents the brightness of the i-th pixel in the first image. ij This represents the brightness of the i-th pixel in the j-th image. Based on the above conditions, overexposed areas are identified, and their weights are set to 0, meaning they are not included in the fusion process. The final fusion result is as follows:

[0118] ,

[0119] Where C is related to the final synthesized bit depth, assuming outbit represents the desired bit depth of the synthesized image, then C is specifically expressed as:

[0120] C = 2(outbit - bit); the previous text recorded "w(Z ij Let represent the weight of the i-th pixel in the j-th image. ij This refers to W(Z) derived earlier. ij), both represent the weight of the i-th pixel in the j-th image. The j in the j-th image and the j-th exposure have the same meaning; one exposure produces one image. W ij = W(Z ij ).

[0121] This embodiment performs channel-wise tone mapping on a high dynamic range image to obtain a low dynamic range image in the RAW domain, including the following steps:

[0122] The first step is to extract the RGB data from the RAW image separately, then map each data channel using the algorithm described below. Finally, the three channels are combined to obtain the final output image. A RAW data format is shown in the attached image. Figure 3 As shown.

[0123] This embodiment first performs global image processing:

[0124] ,

[0125] ,

[0126] ,

[0127] Where c represents the color adaptation parameter, a represents the brightness adaptation parameter, and L represents the grayscale image of the input RAW image, obtained using linear interpolation. av Represents the mean of the grayscale image, where I r|g|b and These represent the RGB values ​​and the pixel mean of the three channels of the RAW image after white balance.

[0128] This represents the pixel value after weighting the pixel brightness point by point.

[0129] I represents the pixel value after weighting by the average pixel brightness. a_r|g|b This represents the pixel value after global processing. a and c have little impact on the final experimental results, so default values ​​of 0.0 and 0.5 are used respectively.

[0130] This embodiment calculates the relevant parameters:

[0131] During image compression, the overall brightness and contrast of the image need to be adjusted. To improve the adjustable range of brightness, an exponential method is used to obtain the image brightness adjustment parameter f, specifically expressed as:

[0132] f = exp(f'),

[0133] f' represents a brightness adjustment parameter that can be manually entered. The larger the f', the higher the brightness, but the image contrast will also be relatively reduced. During the implementation process, it was found that the effect was best when the value was 4.0.

[0134] The contrast parameter m is obtained using an adaptive method, and the specific calculation formula is as follows:

[0135] m = 0.3 + 0.7 * k 1.4 ,

[0136] ,

[0137] k represents the image key value, L max L represents the maximum value of the grayscale image obtained through interpolation. min L represents the minimum value. av The mean of the entire grayscale image.

[0138] This embodiment performs color level mapping after obtaining the brightness and contrast calculation parameters.

[0139] σ(I a_r|g|b )=(f*I a_r|g|b ) m ,

[0140] ,

[0141] I r|g|b V represents the number of pixels in the input image. r|g|b This represents the output data after color level mapping, σ(I) a_r|g|b ) represents the half-saturation constant, I a_r|g|b This represents the pixel value after global processing, where f is the brightness adjustment parameter and m is the contrast parameter. V r|g|b The images are combined to form a RAW image V in RGGB format. Finally, normalization is performed to obtain the globally mapped RAW image. The normalization method is as follows:

[0142] ,

[0143] Where V max V represents the maximum value of the grayscale image V in the output image. min V represents the minimum value. out This indicates the output RAW image.

[0144] Example 3

[0145] This embodiment provides an electronic device, including a processor and a memory, wherein the memory stores computer-readable instructions, which, when executed by the processor, implement a high dynamic range image synthesis and tone mapping method.

[0146] In summary, this invention solves the problem that synthesizing and mapping high dynamic range images on RGB consumes a large amount of computer resources and seriously affects the color and brightness information of the image. This invention has fast calculation speed, low resource consumption, good visual results, and is suitable for hardware-integrated high dynamic range image synthesis, and has a strong market application prospect.

[0147] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A high dynamic range image synthesis and tone mapping method, characterized by, The method comprises the following steps: S1, acquiring multiple RAW images with different exposure degrees in the same scene through a camera; S2, performing exposure fusion on the RAW images with different exposure degrees through a square weighted function to obtain a high dynamic range image synthesized in a RAW domain; and S3, performing channel division tone mapping on the high dynamic range image to obtain a low dynamic range image in the RAW domain; The low dynamic range image in the RAW domain is obtained by extracting rgb three types of data in the RAW image, then performing mapping through an algorithm, and finally combining the three channels of data to obtain the final output image. The high dynamic range image is processed in a channel division manner, including: extracting rgb three types of data in the RAW image, ， ， ， wherein c represents a color adaptive parameter, a represents a brightness adaptive parameter, L represents a gray scale image of the input RAW image, which is obtained by linear interpolation, Lav represents the mean value of the gray scale image, wherein I r|g|b with respectively represent the rgb values and the pixel mean values of the three channels after the RAW image is processed by channel division after white balance, represents the pixel value after the pixel brightness is weighted point by point, represents the pixel value after the average pixel brightness is weighted, a_r|g|b represents the pixel value after global processing; Splitting a high dynamic range image into channels and tone mapping the channels to obtain a low dynamic range image in the RAW domain, comprising: σ(I a_r|g|b ) = (f*I a_r|g|b ) m ， I r|g|b represents a pixel of the input image, V r|g|b represents output data after tone mapping, σ(I a_r|g|b ) represents a half-saturation constant, I a_r|g|b represents a pixel value after global processing, f is a luminance adjustment parameter, and m is a contrast parameter; V r|g|b is combined to form a RAW image V in the RGGB format, and finally normalized to obtain a RAW image after global mapping, in the following manner: ， where V max represents the maximum value of the output image Vgray image, V min represents the minimum value, V out represents the output RAW image.

2. The high dynamic range image synthesis and tone mapping method of claim 1, wherein, In the method, the RAW image is original data converted from light source signals captured by a CMOS or CCD image sensor of the camera into digital signals.

3. The high dynamic range image synthesis and tone mapping method of claim 1, wherein, In the method, the well-exposed area is given a higher weight value when fused, and then a weighted synthesis is performed, and the irradiance value of the image and the image pixel value satisfy the following formula: Z ij = f (E i * t j ) where subscript i represents the i-th pixel, subscript j represents the j-th image, t j represents the exposure time of the j-th image, z ij represents the true pixel value of the i-th pixel in the j-th image, Ei represents the irradiance value of the i-th pixel; the meaning of "f" is a camera response function (Camera Response Function) which is a nonlinear mapping in the camera imaging process.

4. The high dynamic range image synthesis and tone mapping method of claim 1, wherein, In the method, when the RAW format image is synthesized, the camera response function f is expressed as: f = g, where g represents a constant, and the irradiance value E of the image can be expressed as: E = g * I ij may be expressed as: ， t j denotes the exposure time of the jth image, Z ij denotes the pixel value of the ith pixel in the jth image; After obtaining the irradiance value of the image, the image is synthesized, and a quadratic function with a rapid gradient rise in a small pixel position and a rapid gradient drop in a large pixel position is adopted, and the quadratic function is specifically as follows: ， where bit represents the bit number of the input image, Z ij represents the pixel value of the i-th pixel in the j-th image, w(Z ij ) represents the weight value of the i-th pixel in the j-th image.

5. The high dynamic range image synthesis and tone mapping method of claim 4, wherein, In the method, the exposure time ratio Δt is assumed to be: ， wherein t j denotes the exposure time of the jth image, t1denotes the exposure time of the first image; The luminance L of the image is the maximum value of pixels in an RGGB unit block, and the luminance values in the same unit block are the same, and when the following condition is met, the pixel is considered to be an overexposed pixel: ， where black represents the value of the black level, bit represents the number of bits of the input image, m is an adjustable value, At represents the exposure time ratio, L i1 represents the brightness of the i-th pixel of the first image, L ij represents the brightness of the i-th pixel in the j-th image; according to the above condition, it is judged that the overexposed region is set to 0, that is, it does not participate in fusion, and the final fusion result is as follows: ， Wherein C is related to the final synthesis bit number, assuming that outbit represents the bit number of the expected synthesized image, then C is specifically represented as C=2(outbit-bit).

6. The high dynamic range image synthesis and tone mapping method of claim 1, wherein, In the method, the overall luminance and contrast of the image need to be adjusted in the process of compressing the image, and the luminance adjustment parameter f of the image is obtained in an exponential manner, and is specifically represented as: f=exp(f’), f’ represents a manually input luminance adjustment parameter, and the larger f’ is, the higher the luminance is, and the contrast of the image is also relatively reduced; The contrast parameter m is obtained by an adaptive method, and the specific calculation formula is as follows: m = 0.3 + 0.7*k 1.4 , ， k represents a key value of the image, Lmax represents the maximum value of the gray scale image obtained through interpolation, Lmin represents the minimum value, and Iav represents the average value of the entire gray scale image.

7. An electronic device, comprising: The method comprises a processor and a memory, wherein the memory stores computer readable instructions, and the computer readable instructions are executed by the processor to implement the high dynamic range image synthesis and tone mapping method according to any one of claims 1 to 6.

Images