TW201526640A - High dynamic range image composition apparatus for carrying out exposure mapping based on individual pixel and the method thereof - Google Patents

Abstract

Disclosed is a high dynamic range image composition apparatus for carrying out exposure mapping based on individual pixel. It comprises an image capturing module, an image processing module, and a recursive control module. The image capturing module is used for acquiring a plurality of images having different exposure value respectively. The image processing module is used for performing exposure value tuning means so as to perform exposure synthesis for the pixels based on exposure value between pixels at the same position in images with close exposure values. Accordingly, synthesis image can be composed. The recursive control module is used for re-transmitting generated composite image to the image processing module to be processed until the image processing module only generates a composite image.

Description

High dynamic range image synthesizing device based on individual pixels for exposure mapping and method thereof

The present invention relates to a high dynamic range image synthesizing apparatus and method thereof, and more particularly to a high dynamic range image synthesizing apparatus and method thereof for performing exposure mapping based on individual pixels.

Due to the limitations of the photosensitive element on the hardware, any photographic device has a limited dynamic image range. In a frame of image, if the brightness difference between the bright part and the dark part is too large, in the case of a low dynamic range, it may be possible to make the cut image appear bright and white at the position of the bright part, and the position of the dark part is full. In the case of black, the details of the images (such as outlines, lines, etc.) in these high-light, low-light areas are lost, causing the problem that the images cannot faithfully present the real picture.

In order to overcome the above problems, the industry has developed a variety of image processing methods that can improve the dynamic range of images. The images generated by these image processing methods are generally referred to as High Dynamic Range Imaging (HDRI). In general, the more common high dynamic range imaging technology is to image the same scene based on different exposure values by adjusting the parameters of the photographic equipment, and synthesize these images to form images by image synthesis. The high-gloss area and low-light area in the middle can preserve the details, so as to faithfully present the original image, and improve the dynamic range that can be displayed in the image. This technology helps to perform clear photography in high contrast environments, or for eye protection masks used by welding workers. By welding, high-contrast images of welding are processed for immediate operation during electric welding. The location of the solder joints is found on the display and is used in a wide range of applications.

However, in the image synthesis process described above, the image is synthesized based on the object captured by the lens while the object is being shot, and the object captured in the lens is moved. At the same time, the object between each image 将会 will produce a slight offset (depending on the speed of movement and the number of images taken per second), which makes the image 祯 and the image 略 slightly different, which The offset of the class may be considered as noise in the image synthesis, causing image distortion.

Furthermore, since all pixels in each image are synthesized using only the same exposure value as a reference value, in tone mapping (Tone Mapping), each pixel has its brightness (Quantigraphic Measure) and camera response function (Camera Response). ) and the degree of noise interference, pixels with extremely high or very low brightness may still have some highlights or low-light areas after image synthesis. Because the contrast is too large, the details cannot be clearly expressed, so that even in the image After the high dynamic range processing, there are still some details that cannot be represented, and the complete tone mapping cannot be achieved, which is necessary for improvement.

The object of the present invention is to solve the problem that when a high dynamic range image is synthesized, noise is generated due to delay in photographing, and tone mapping may result in loss of details of bright or dark regions in the image.

To solve the above problems, the present invention provides a high dynamic range image synthesizing device based on an individual pixel exposure mapping, comprising: an image capturing module, an image processing module, and a recursive control module. The image capture module is configured to obtain a plurality of images having different exposure values. The image processing module is configured to perform exposure synthesis on the pixels according to the exposure values between the pixels at the same position in the images with similar exposure values, so as to form a composite image. The recursive control module retransmits the generated composite image to the image processing module for processing until the image processing module generates only one composite image.

Further, the exposure value adjustment means assigns the exposure value of the pixel at the same position in the image with similar exposure values to each other according to the following formula:

Further, the exposure synthesis system generates a synthesized pixel according to the exposure value after the weight distribution in the exposure value adjustment means, and substitutes the following formula:

Further, the high dynamic range image synthesizing device based on the full pixel exposure value mapping further includes an extended synthesis module, wherein if the difference between the exposure values of the Nth pixels in the two images is greater than a threshold, then The extended synthesis module generates a virtual pixel, and generates a first modified pixel and a second modified pixel according to the difference between the Nth pixel and the virtual pixel in the two images, and then according to the first The synthesized pixel is generated by a corrected pixel and a difference in exposure value between the second modified pixels.

Further, the exposure value of the virtual pixel is an average value of exposure values of the Nth pixel in the two images.

Further, the image processing module generates a corresponding plurality of mapping functions according to each of the exposure value differences, and generates a look-up table (LUT) according to the mapping function, and the image processing module The group is substituted into the lookup table according to the difference in exposure values between the Nth pixels in the two images to generate the synthesized pixels.

Another object of the present invention is to provide a high dynamic range image synthesis method based on full pixel exposure value mapping, the method comprising the steps of: (a) obtaining a plurality of images having different exposure values; (b) according to two Exposing the pixels to the pixels at the same position in the image with the same exposure value by an exposure value adjustment method; (c) determining that if the number of synthesized images is not 1, the The composite image is passed back to step (b) and processed again.

Further, the exposure value adjustment means generates the synthesized pixel according to the following formula:

Further, the pixel synthesis program includes the steps of: generating a composite exposure value via a weight assignment program; and generating the synthesized exposure value according to the A synthetic pixel.

Further, the weight distribution program includes the steps of: measuring a first exposure value of the Nth pixel in the image with lower exposure value, and a first weight coefficient; measuring one of the Nth pixels in the image with higher exposure value a second exposure value, and a second weight coefficient; and multiplying the first exposure value by the first weight coefficient and multiplying the second exposure value by the second weight coefficient, adding, dividing by The sum of the first weighting factor and the second weighting factor to generate the composite exposure value, or can be written as follows:

Further, the first weight coefficient and the second weight coefficient are inverse functions of the luminance dynamic range of the Nth pixel in the two images by Shannon's entropy, or may be written as follows:

Further, the weight distribution program further includes the following steps: if the difference between the exposure values of the Nth pixels in the two images is greater than a threshold, a virtual pixel is generated, and the exposure value of the virtual pixel is between the two Between the Nth pixels in the image; performing the weight distribution process on the virtual pixel and the Nth pixel having lower exposure values in the two images to generate a first modified pixel; comparing the exposure value between the virtual pixel and the two images The high Nth pixel performs the weight assignment process to generate a second modified pixel; and replaces the N pixel in the image with the lower exposure value with the first modified pixel, and replaces the higher exposure value with the second modified pixel The Nth pixel in the image and the weight assignment procedure is executed.

Further, the pixel synthesis program further includes the following steps: generating a corresponding mapping function according to each of the exposure value differences, and generating a look-up table (LUT) according to the mapping function. And substituting the lookup table according to the difference in exposure values between the Nth pixels in the two images to generate the synthesized pixel.

Therefore, the present invention has the following beneficial effects over the above prior art:

The high dynamic range image synthesizing device and the method thereof based on the full pixel exposure value mapping of the present invention can completely display the image details located in the extremely bright and extremely dark portions by synthesizing each pixel in the image separately.

2. The extended synthesis module of the present invention can generate a virtual pixel with an exposure value centered when the difference between the exposure values of the pixels in the two images is too high, and synthesize the virtual pixel and the two images, thereby reducing the height. The image resulting from the difference in exposure value is jagged, or the synthetic pixel is beyond the critical point of the dynamic range.

3. The present invention further provides a method for generating a lookup table, thereby accelerating the efficiency of the overall algorithm of the present invention, and can be applied to a dynamic image such as a welding picture for high dynamic range image synthesis processing without affecting the smoothness of the picture.

100‧‧‧Image synthesizer

10‧‧‧Image capture module

20‧‧‧Processing unit

21‧‧‧Image Processing Module

22‧‧‧Return control module

23‧‧‧Extended Synthetic Module

Steps S110 to S160‧‧

1 is a block diagram of a high dynamic range image synthesizing apparatus based on full pixel exposure value mapping according to the present invention.

2 is a schematic diagram of image synthesis of the present invention.

3-1 and 3-2 are schematic diagrams of the image synthesis output of the present invention.

4 is a schematic diagram of a lookup table of the present invention.

5 to 6 are flow charts of the method of the present invention.

For the structural features and operation methods of this case, and with the illustrations, please refer to it later.

For the technology of the present invention, please refer to FIG. 1 , which is a block diagram of a high dynamic range image synthesizing device for performing exposure mapping based on individual pixels according to the present invention. As shown in the figure, the image synthesizing device 100 includes: an image. The module 10 and the processing unit 20 are captured. The image capturing module 10 is selected from the group consisting of a camera lens, a camera lens, and a scanner lens for capturing a scene. The image capturing module 10 captures the same scene based on different exposure values, thereby obtaining a plurality of images having different exposure values. The processing unit 20 includes an image processing module 21 and a recursive control module 21. This treatment The unit 20 can be an image processing unit in an imaging device (such as a camera, a camera, a scanner, etc.) or a Field Programmable Gate Array (FPGA) installed in the camera device. ) The program in the chip. The processing unit 20 can be an arithmetic unit in the device, and the image processing module 21 and the recursive control module 22 can be a drawing software or The program in the image processing software is not limited in the present invention.

The term "individual pixel exposure mapping" of the present invention is as follows: in the process of synthesizing high dynamic range images, each image contains N pixels, and when the two images are combined, the same in the two images In the position, there are corresponding pixels (Pixel). For convenience of subsequent description, the pixels located at the same position are called Pixel Pairs, and the present invention is based on the difference of exposure values between each pair of pixels. The value, Weighted Sum Method, performs high dynamic range image synthesis, which is different from the previous method in which all pixels in the entire image use the same exposure value. The pixels obtained by the individual pixel exposure mapping are generated according to the following formula (1):

In formula (1), both ^ q and ^ k correspond to a coordinate in the vertical and horizontal positions in the image, that is, the coordinates of the pixel on the image, which is a simplified representation formula, where ^ q and ^ k The coordinates are omitted. Wherein, P _mapped value of the pixel that is mapped treatment; F imaging device to the camera response function (camera response); ^ q representing the pixel whichever camera function in response to the return function, and then divided by the known exposure, etc. The obtained brightness estimation value; ^ k is the inverse function of the pixel value of the pixel after taking the inverse function of the inverse camera response function, and then dividing by the image into the light to obtain the exposure level estimation value, the exposure level ^ k and The relationship between the exposure values EV of the pixels is ^ k = 2 ^EV , and the main object of the present invention is to combine the exposure values EV of the respective pixels in the two images to obtain a more appropriate exposure level ^ k .

The image processing program executed by the present invention will be described in detail below.

For the purpose of understanding the image synthesizing method of the present invention, please refer to FIG. 2, which is a schematic diagram of image synthesis according to the present invention. As shown in the figure, the image processing module 21 synthesizes images with similar exposure values. The total number of images is I, and the image processing module 21 performs synthesis. After the program, I-1 composite images will be generated, and the recursive control module 22 will retransmit the generated composite images to the image processing module 21 for processing. After I-1 recursive, finally only A composite image is retained, which is a high dynamic range image that retains all details after dynamic range compression.

More specifically, the image processing module 21 selects two images with similar exposure values in the plurality of images acquired by the image capturing module 10. For convenience of description, the two images are referred to as images i and images. j, wherein the overall exposure value of the image i is smaller than the overall exposure value of the image j, and the Nth pixel in the image i is referred to as a pixel P _Ni , and the Nth pixel in the image j is referred to as a pixel P _{Nj .} The pixel P _Ni and the pixel P _Nj are a pair of pixel pairs located at the same position. According to the foregoing steps, the image processing module 21 calculates an exposure value difference ΔEV _i,j between the pixel P _Ni and the pixel P _Nj , and according to the exposure value difference ΔEV _i,j and the pixel P _Ni The exposure value of the pixel P _Nj is weighted to generate a composite exposure value EV _i,j , and the synthesized exposure value EV _{i,j is} substituted into the above formula (1) to obtain the synthesized pixel generated by the mapping. After the image processing module 21 synthesizes the N pixel pairs, N synthesized pixels are generated, and the synthesized pixels can be combined into a high dynamic range image.

The term "weight distribution" of the present invention distributes the exposure values of two pixels (P _Ni , P _Nj ) according to a certain specific gravity, and the distribution method is determined by the following formula (2):

In formula (2), the exposure value after weight assignment is EV _i,j ; and EV _i , EV _i represent the exposure values of the two pixels themselves, and C _i and C _j are the weight coefficients for determining the two pixels. And the resulting exposure value is EV _i,j will be closer to the exposure value of the higher weight coefficient.

However, the weight coefficients C _i and C _j must actually depend on the variation range of the brightness of the pixels. Due to the interference of the noise, a pixel value may be generated when mapping the pixel P _Ni and the pixel P _Nj . The problem of diverging maps to multiple pixel values, thus causing uncertainty on the map. In this regard, the processing method adopted by the present invention is to obtain the dynamic range of the brightness of the pixels, that is, the maximum value of the brightness of the pixel is divided by the minimum value and then take the logarithm thereof, and according to the information theory, the summer entropy value (Shannon's The model of entropy) obtains the uncertainty equation of the dynamic range in the image, which is selected from the sample space generated by the exposure of the pixel and the vector composed of the difference of the exposure values of the image pixels having different exposure values. In the case of simplifying the camera response function of the camera device, the inverse function of the Xia's entropy value is obtained, and the respective deterministic functions of the two images can be obtained, as in the formula (3), taking the pixel P _Ni as an example:

Wherein, C is the weight coefficient; Y _i is the inverse value of the pixel value of the pixel P _Ni as the inverse of the camera response function, and divided by the brightness and exposure level, in the mathematical sense, Y _i Is the ratio of the exposure level estimate ^ k _i obtained by the camera response function back to the actual exposure level of the pixel, or writing:

H ^N refers to the entropy value of the normal distribution of Y _i in the sample space of the vector ζ (defined as [brightness q, exposure value difference ΔEV _{i, j} ]) under the normal distribution. After substituting the pixel values of the two pixels to be synthesized into the deterministic function, the weight coefficient of the pixel at the time of weight distribution can be obtained, and the exposure value which is easier to display the details can be synthesized according to the weight coefficient. For example, for a darker block, the weight coefficient of the pixel with lower exposure value is larger than that of the pixel with high exposure value, and the higher the difference between the exposure values between the two pixels, the mutual The difference between the weight coefficients will also increase. However, the above methods are for illustrative purposes only. Different deterministic equations can be derived from other mathematical models depending on the camera response function. For details, please refer to the paper Technology and Society (ISTAS), 2013 IEEE International Symposium on 27-29 June 2013 , pages 98-106, High Dynamic Range Tone Mapping Based on Per-Pixel Exposure Mapping, while the present invention only discusses the condition that the camera response function is a logarithmic function, and the calculation method is not intended to be limited by the present invention.

As described above, when the difference between the exposure values of the Nth pixel pairs in the two images is too high, a large difference is generated between the weight coefficients of the pixels, and the result may result in a high dynamic range image. Still unable to display complete image details. Please refer to FIG. 3-1 and FIG. 3-2 together, which is a schematic diagram of the image synthesis output of the present invention, as shown in the figure: FIG. 3-1 and FIG. 3-2 respectively represent a one-dimensional image, and its horizontal axis. The representative image contains 600 pixels, and the vertical axis indicates the pixel values of the pixels. As shown in Figure 3-1, when the difference in exposure value is too high, only the weight distribution is used to perform pixel synthesis, which will result in pixel values. Continuous, so that the image may produce a jagged phenomenon, which is visible in the pixels of the last two arc segments of Figure 3-2, a darker pixel P _i and an extremely bright pixel P _j The synthesized pixels may still exceed the threshold of the dynamic range that can be displayed.

In order to solve the above problem, the high dynamic range image synthesizing device 100 based on the individual pixel exposure mapping further includes an extended synthesis module 23, if the Nth pixel (pixel P _Ni , pixel P _Nj ) in the two images When the difference between the exposure values is greater than a threshold, the extended synthesis module 23 generates a virtual pixel, and the pixels P _Ni and P _Nj are respectively weighted with the virtual pixel to obtain a first modified pixel. And a second modified pixel. At this time, if the difference between the exposure value between the first modified pixel and the second modified pixel has been reduced to a reasonable range, the synthesized pixel may be directly generated according to the difference of the exposure value. If the difference between the exposure values of the first modified pixel and the second modified pixel is still too high, the extended synthesis module 23 generates the virtual pixel again and repeats the above steps until the exposure value difference enters a reasonable range. There is no specific value for the threshold, and the user depends on the detailed presentation requirements of the high dynamic range image. If the threshold is lowered, the image with higher quality is obtained, but if the threshold is too low, On the contrary, it may cause the image to be blurred, so the threshold value must be designed within a reasonable range according to the user's needs. In a preferred embodiment, the exposure value of the virtual pixel is preferably an average value of the exposure values of the Nth pixel in the two images, so that the generated correction pixel can minimize the difference in the exposure value. The image processed by the program is equivalent to the extension of the weight distribution. As can be seen from FIG. 3-1, the pixel values generated by the synthetic pixels processed by the extended synthesis module 23 are relatively smooth and have no jaggedness. Moreover, it can be seen from 3-2 that the pixels generated by the extension synthesis module 23 can be effectively limited to the dynamic range without causing problems such as loss of details of the image due to excessive difference in exposure values.

If the image composition method based on the weight distribution is executed by an algorithm and the number of images acquired by the image capturing module is I, the time complexity of the algorithm is O(I ² ), that is, The more the number of images obtained, the longer it takes to perform the calculation; on the other hand, if the difference between the exposure values is too high, the extended synthesis module needs to generate V virtual pixels, the overall time of the algorithm is complicated. The degree is further degraded to O((VI) ² ). If the dynamic image (such as the welding image mentioned in the prior art) is processed, the frame number of the whole frame (FPS) will be made. It is too low to display images smoothly. Therefore, the present invention further provides a method for accelerating the overall algorithm. The image processing module 21 generates a corresponding mapping function according to each of the exposure value differences, and generates a lookup table according to the mapping function. The look-up table (LUT), and the image processing module 21 substitutes the difference between the exposure values of the Nth pixels in the two images into the lookup table to generate the synthesized pixels. Please refer to FIG. 4 , which is a schematic diagram of a lookup table of the present invention. Specifically, the lookup table is generated by separately inserting a pixel pair having a higher difference in exposure value into the virtual pixel through the extension synthesis module 23, and then synthesizing each time the weight distribution of the pixel pair is performed. The mapping result produced by the difference of the exposure values is summarized into a mapping function, and finally all the mapping functions are combined on the same plane to complete the lookup table of FIG. In actual use, the processing unit only needs to bring the pixel value and the exposure value difference of each pixel pair in the image into the lookup table to generate the synthesized pixel, and in an ideal state, the overall algorithm can be generated. The time complexity is optimized to O(1), which in turn reduces the synthesis processing time for high dynamic range images. Since the manner in which the mapping function is obtained is summarized as a technique familiar to those of ordinary skill in the art, detailed steps are omitted in the present invention.

The following describes the high dynamic range image synthesis method based on the full pixel exposure value mapping of the present invention. Referring to FIG. 5 to FIG. 6 , it is a flowchart of the method of the present invention. As shown in the figure, the method includes the following steps: obtaining a complex number Each of the images has images of different exposure values, and the image contains N pixels (S110). Secondly, according to the N-th pixel of the same position in the image with similar exposure values, the pixels are subjected to exposure synthesis by an exposure value adjustment method to form a composite image (S120), wherein the exposure value adjustment means The method includes the steps of: generating a composite exposure value via a weight assignment process (S121), and generating a composite pixel based on the composite exposure value (S122). Next, the generated N synthesized pixels are combined into a composite image (S130). Finally, it is judged that if the number of synthesized images is not 1, the synthesized image is returned to step S120 and processed again (S140). By the above method steps, the present invention can synthesize each pixel pair at the same position in a plurality of images, and finally obtain a high dynamic range image.

The detailed process of pixel synthesis of the present invention is further described below. The weight distribution procedure is as follows: in the same location pixel, a first exposure in an image with a lower exposure value is measured. Value, and a first weighting factor. A second exposure value in the image with a higher exposure value is measured, and a second weight coefficient is measured. Finally, multiplying the first exposure value by the first weight coefficient and multiplying the second exposure value by the second weight coefficient, adding, dividing by the first weight coefficient and the second weight coefficient And generating the synthesized exposure value and generating a synthesized pixel according to the synthesized exposure value. The first weight coefficient and the second weight coefficient are inverse functions of the luminance dynamic range of the Nth pixel in the two images by Shannon's entropy.

In order to avoid problems such as image discontinuity and loss of detail due to excessive difference in exposure value, the weight distribution program further includes the following steps: if the difference between the exposure values of the Nth pixels in the two images is greater than one In the threshold, a virtual pixel is generated, and the exposure value of the virtual pixel is between the Nth pixels in the two images. In a preferred embodiment, the exposure value of the virtual pixel is the The average of the exposure values of the N pixels. Then, the weight distribution process is performed on the virtual pixel and the Nth pixel with lower exposure values in the two images to generate a first modified pixel, and the virtual pixel and the Nth pixel having higher exposure values in the two images are executed. The weight is assigned to generate a second corrected pixel. Finally, the Nth pixel in the image with the lower exposure value is replaced by the first modified pixel, and the Nth pixel in the image with the higher exposure value is replaced by the second modified pixel, and the weight distribution process is executed.

Finally, in order to improve the efficiency of the overall algorithm of the method of the present invention, the pixel synthesis program further includes the following steps: generating a corresponding mapping function according to each of the exposure value differences, and generating according to the mapping function; A look-up table (LUT). And substituting the lookup table according to the difference between the exposure values between the Nth pixels in the two images to generate the synthesized pixel.

In summary, the high dynamic range image synthesizing device based on the full pixel exposure value mapping of the present invention and the operating method thereof can obtain a plurality of images having different exposure values through the image capturing module, and then pass the image processing mode. The combination becomes a composite image with higher detail rendering, and is returned to the control module to finally generate a high dynamic range image. In addition, when the difference between the exposure values of the pixels in the two images is too high, the virtual pixels can be inserted through the extended synthesis module, thereby avoiding the problem that the final image is discontinuous or the details are lost. The method of the present invention can accelerate the operation of the overall algorithm by establishing a lookup table, so that the high dynamic range image synthesizing device based on the full pixel exposure value mapping of the present invention and the operation method thereof can also perform high dynamic range imaging for continuously changing images. Synthetic processing without reducing the fluency of its picture.

The present invention has been described in more detail with reference to the preferred embodiments described above, but the present invention is not limited to the embodiments described above, and various changes and modifications may be made to the structures within the scope of the technical idea disclosed herein. Such changes and modifications are still within the scope of the invention.

100‧‧‧Image synthesizer

10‧‧‧Image capture module

20‧‧‧Processing unit

21‧‧‧Image Processing Module

22‧‧‧Return control module

23‧‧‧Extended Synthetic Module

Claims (14)

A high dynamic range image synthesizing device based on an individual pixel exposure mapping, comprising: an image capturing module for acquiring a plurality of images respectively having different exposure values; and an image processing module for using the two exposure values Exposure values between pixels in the same position in the similar image are subjected to exposure value adjustment means for combining and synthesizing the pixels to form a composite image; and a recursive control module retransmits the generated synthesized image to the The image processing module processes until the image processing module generates only one composite image. The image synthesizing device according to claim 1, wherein the exposure value adjusting means assigns the exposure value of the pixel at the same position in the image with similar exposure values to each other according to the following formula: The image synthesizing device according to claim 1, wherein the exposure synthesis system generates a synthesized pixel according to the exposure value after the weight distribution in the exposure value adjustment means, and substitutes the following formula: The image synthesizing device of claim 3, further comprising an extended synthesis module, wherein a virtual pixel is generated when the difference between the exposure values of the pixels at the same position in the image is greater than a threshold value, and according to the second The difference between the pixels at the same position and the difference between the virtual pixels in the image respectively generates a first modified pixel and a second modified pixel, and then according to the first modified pixel and the second modified pixel. The difference in exposure value produces the synthesized pixel. The image synthesizing device of claim 4, wherein the exposure value of the virtual pixel is an average of exposure values of pixels at the same position in the two images. The image synthesizing device according to any one of claims 1 to 3, wherein the image processing module generates a corresponding plurality of mapping functions according to an exposure value adjustment means, and generates a mapping function according to the mapping function. Look-up Table (LUT), the image processing module The lookup table is substituted into the lookup table according to the difference in exposure values between pixels in the same position in the image to generate the synthesized pixel. A high dynamic range image synthesis method for performing exposure mapping based on individual pixels, the method comprising the steps of: (a) obtaining a plurality of images having different exposure values; (b) pixels of the same position in images having similar exposure values Performing exposure synthesis on the pixels by an exposure value adjustment method to form a composite image; and (c) determining that if the number of synthesized images is not 1, the composite image is transmitted back to Step (b) is processed again. The method of claim 5, wherein the exposure value adjustment means comprises the steps of: generating a composite exposure value via a weight assignment procedure; and generating the composite pixel based on the composite exposure value. The method of claim 8, wherein the exposure value adjustment means generates the synthesized pixel according to the following formula: The method of claim 8, wherein the weighting process comprises the steps of: measuring a first exposure value of the pixel in the image with a lower exposure value, and a first a weight coefficient; in the same position pixel, measuring a second exposure value of the pixel in the image with higher exposure value, and a second weight coefficient; and multiplying the first exposure value by the first weight coefficient and The second exposure value is multiplied by the second weight coefficient and added by the sum of the first weight coefficient and the second weight coefficient to generate the composite exposure value, or can be written as follows: The method of claim 10, wherein the first weight coefficient and the second weight coefficient are the inverse of the brightness dynamic range of the same position pixel in the two images by Shannon's entropy Function, or can be written as follows: The method of claim 10, further comprising the step of: generating a virtual pixel if the difference in exposure values between pixels in the same location in the image is greater than a threshold, the virtual pixel The exposure value of the pixel is between two exposure values of the same position pixel in the image; and in the same position pixel, the weight distribution process is performed on the virtual pixel and the pixel with lower exposure value in the image. Generating a first modified pixel, performing the weight assignment process on the virtual pixel and the pixel with higher exposure value in the two images to generate a second modified pixel; and replacing the image with the lower exposure value by the first modified pixel a pixel, and replacing the pixel in the image with the higher exposure value with the second modified pixel, and executing the weight assignment procedure. The method of claim 12, wherein the exposure value of the virtual pixel is an average of exposure values of pixels at the same position in the image. The method of claim 7, wherein the pixel synthesis program further comprises the steps of: generating a corresponding mapping function according to each of the exposure value differences, and generating a mapping according to the mapping function; a look-up table (LUT); and substituting the look-up table according to the difference in exposure values between pixels at the same position in the image to generate the synthesized pixel.