Dual-gain HDR synthesis method and system for sCMOS sensors

Abstract

The present invention provides a dual-gain HDR synthesis method and system for an sCMOS sensor, relating to the field of image processing. The method utilizes an sCMOS sensor to simultaneously acquire two images, one high-gain and one low-gain; performs a linear grayscale value analysis on the high-gain and low-gain images; performs image histogram statistics to determine gain and exposure time settings; and performs grayscale stretching on the high-gain and low-gain images to synthesize an HDR image based on the grayscale value relationship. By performing the grayscale value linear relationship analysis and grayscale stretching on the high-gain and low-gain images, a high dynamic range image can be rapidly synthesized, eliminating image registration issues and simplifying the HDR image synthesis process. By performing the grayscale value linear relationship analysis and grayscale stretching on the high-gain and low-gain images, a high dynamic range image can be rapidly synthesized.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a dual-gain HDR synthesis method and system for an sCMOS sensor. Background Art

[0002] High dynamic range (HDR) imaging technology plays an increasingly important role in modern image processing and computer vision. The primary goal of HDR imaging is to overcome the dynamic range limitations of traditional imaging systems, enabling the simultaneous capture and display of details in both the brightest and darkest parts of a scene. Traditional imaging sensors often struggle to capture both highlights and shadows in a scene in a single exposure, resulting in images that are either overexposed (no details in highlights) or underexposed (no details in shadows). To address this issue, HDR imaging technology was developed.

[0003] Traditional HDR imaging methods are typically based on multiple exposure techniques, which involve taking multiple exposures of the same scene, obtaining multiple images at different exposure times, and then combining these images into a single HDR image. While this method is effective, it has several significant drawbacks:

[0004] (1) Image registration is difficult: Since multiple exposures take time, moving objects in the scene can cause image inconsistencies, increasing the complexity of image registration and making it prone to artifacts such as ghosting.

[0005] (2) High processing complexity: The multiple exposure method requires complex image processing algorithms to align and synthesize images, which increases the computational cost and processing time.

[0006] (3) Lack of real-time performance: Multiple exposures require multiple shots, making it difficult to achieve real-time HDR imaging, which limits its application in dynamic scenes.

[0007] The dual-gain feature of sCMOS sensors offers a new solution to the image registration challenges faced by traditional multi-exposure HDR synthesis methods. sCMOS sensors can simultaneously output high-gain and low-gain images, which, when synthesized using appropriate algorithms, can create images with a wide dynamic range. Summary of the Invention

[0008] In view of this, to solve the image registration problem in traditional multiple exposure methods and simplify the HDR image synthesis process, this paper proposes a dual-gain HDR synthesis method and system for sCMOS sensors. By performing grayscale linear relationship analysis and grayscale stretching on high-gain and low-gain images, high dynamic range images can be quickly synthesized. This aims to eliminate image registration problems and simplify the HDR image synthesis process. By performing grayscale linear relationship analysis and grayscale stretching on high-gain and low-gain images, high dynamic range images can be quickly synthesized.

[0009] The present invention is achieved by adopting the following technical solutions:

[0010] In a first aspect, the present invention provides a dual-gain HDR synthesis method for an sCMOS sensor, the method comprising the following steps:

[0011] Use sCMOS sensor to simultaneously acquire high-gain (GH) and low-gain (GL) images;

[0012] Perform linear relationship analysis on grayscale values ​​of high-gain and low-gain images;

[0013] Perform image histogram statistics to determine gain and exposure time settings;

[0014] Grayscale stretching is performed on high-gain and low-gain images, and HDR images are synthesized based on the grayscale value relationship.

[0015] As a further solution of the present invention, when a sCMOS sensor is used to simultaneously acquire a high-gain image and a low-gain image, the high-gain image is used to capture details of darker parts of the scene, and the low-gain image is used to capture details of brighter parts of the scene.

[0016] As a further solution of the present invention, the sCMOS sensor is internally amplified by dual gain channels and then quantized and output by an ADC converter, and an image with a higher dynamic range is synthesized off-chip.

[0017] As a further solution of the present invention, in an sCMOS sensor, if the voltage amplitude corresponding to the full well charge is V m , then the output voltage range after the analog-to-digital converter is expressed as [0, V m ], that is, the signal voltage entering the column bus is V∈[0,V m ], define the gain channel magnification as G, KADC (V / DN) as the ADC converter unit, then the expression of the output data D after a bit precision quantization is:

[0018]

[0019] When V is at its minimum value, i.e., the system noise, in order to quantize V into a digital value for reading, at least At this time, the larger the value of gain G is, the smaller the signal data is read out.

[0020] As a further solution of the present invention, when the pixels are not saturated, the grayscale value relationship of the corresponding points of the high-gain image and the low-gain image is analyzed. When the pixels are not saturated, the grayscale values ​​output by the two gain channels are approximately equal to the set gain values.

[0021] As a further embodiment of the present invention, in an sCMOS sensor, when dual-channel image reading is used, the dynamic range of readable data is defined as:

[0022]

[0023] Where, max(D L ) represents the maximum quantization value of the low gain channel output, the maximum quantization value is 2 a -1, min(D H ) represents the minimum data that can be quantified by the high-gain channel. The minimum data that can be quantified by the high-gain channel is 0. The dynamic range of the image data read by the sensor with dual-gain channel design is improved by 20*log(G H / G L ), the HDR image is obtained by synthesizing the LDR images output by the two channels.

[0024] As a further solution of the present invention, when performing image histogram statistics and determining the gain and exposure time settings, if the grayscale range of a bit is evenly divided into b grayscale levels, where a and b are both even numbers, then the number of dark pixels (grayscale value 0) is defined as (S low ), saturated pixel (gray value is 2 a -1) (S high ), calculate the number of pixels in the upper and lower parts of the histogram of the two images as:

[0025]

[0026] Where P is the percentage of pixels in the histogram, s represents low gain, l represents high gain, H represents the image histogram, h represents the grayscale level (1-b), and N represents the number of all pixels. t,s ≈80%, P t,l ≈80%, then the gain and exposure time are adjusted; P t，s Indicates the percentage of pixels in the low-gain histogram; H t，s (h) represents the number of pixels of gray level (1-b) in the low gain histogram; P t，l Indicates the percentage of pixels in the high gain histogram; H t，l(h) represents the number of pixels at gray level (1-b) in the high gain histogram.

[0027] As a further solution of the present invention, when the pixels are not saturated, the grayscale value relationship between the corresponding points of the high-gain image and the low-gain image is analyzed as follows:

[0028] H x，y =k*L x，y +b H x , y<2 a -1

[0029] Where H x,y Represents the grayscale value of high-gain image, L x,y Represents the grayscale value of the low-gain image, k and b represent coefficients. Each set of gain settings will result in a fixed set of k and b; x and y represent the pixel coordinates.

[0030] As a further solution of the present invention, grayscale stretching is performed on high-gain and low-gain images. When synthesizing an HDR image based on the grayscale value relationship, the idea of ​​directly synthesizing the HDR image with useful information in the high- and low-gain images is adopted. In combination with the high- and low-gain grayscale characteristics, when the following formula is satisfied:

[0031]

[0032] Where a L Indicates the segmentation point of the high-gain image histogram; a H Indicates the segmentation point of the high-gain image histogram;

[0033] Use linear stretching to x,y ≤a L and L x,y ≥a L The area is stretched so that the grayscale range of the fused image is distributed in the entire [0, S], then:

[0034]

[0035] Where H′ and L′ represent the images of H and L after a certain grayscale linear transformation respectively;

[0036] The synthesized image GHDR is represented as:

[0037]

[0038] In a second aspect, the present invention further provides a dual-gain HDR synthesis system for an sCMOS sensor, for executing a dual-gain HDR synthesis method for an sCMOS sensor, the dual-gain HDR synthesis system comprising:

[0039] Dual-gain image acquisition module, used to simultaneously acquire high-gain and low-gain images using an sCMOS sensor;

[0040] Gray value linear relationship analysis module, used to perform gray value linear relationship analysis on high gain and low gain images;

[0041] Image histogram statistics module, used to perform image histogram statistics and determine gain and exposure time settings;

[0042] Grayscale stretching module, used to perform grayscale stretching processing on high-gain and low-gain images;

[0043] The HDR image synthesis module is used to synthesize HDR images based on the grayscale value relationship.

[0044] The present invention also includes a computer device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the dual-gain HDR synthesis method for the sCMOS sensor.

[0045] The present invention also includes a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the dual-gain HDR synthesis method for the sCMOS sensor.

[0046] Compared with the existing technology, the dual-gain HDR synthesis method and system for sCMOS sensors provided by the present invention address the limitations of traditional HDR imaging methods, especially the difficulties in image registration, high processing complexity, and lack of real-time performance in multiple exposure methods. It proposes an efficient and accurate HDR image generation solution with the following beneficial effects:

[0047] 1. Avoids image registration issues. This method leverages the dual-gain characteristics of the sCMOS sensor to simultaneously capture high-gain and low-gain images, fundamentally avoiding image registration issues caused by multiple exposures. Because the two images are acquired within the same frame time, image inconsistencies caused by object movement in the scene are eliminated, and ghosting and artifacts caused by registration errors are eliminated.

[0048] 2. Improved dynamic range. By analyzing the grayscale value relationship between high-gain and low-gain images and performing corresponding grayscale stretching and synthesis, the system can achieve high dynamic range imaging with a single exposure. The high-gain image captures low-brightness details, while the low-gain image preserves high-brightness details. The combination of the two covers a wider brightness range, allowing the image to simultaneously display details in both dark and bright areas of the scene.

[0049] 3. Simplified processing. Compared with traditional multiple-exposure methods, this system's approach eliminates the need for complex image registration and alignment, simplifying the HDR image synthesis process. Through precise grayscale linear relationship analysis and grayscale stretching, high-gain and low-gain image fusion can be quickly and efficiently completed, reducing computational complexity and improving processing efficiency.

[0050] 4. Enhanced real-time performance. Because sCMOS sensors can simultaneously output high-gain and low-gain images within a single frame, this method supports real-time HDR imaging. This is particularly important for applications in dynamic scenes, such as video surveillance, real-time scientific experiment imaging, and real-time medical imaging, significantly improving system response speed and imaging quality.

[0051] 5. Improved image quality. This invention effectively enhances image contrast and detail through grayscale stretching and synthesis strategies. The combination of high-gain and low-gain images not only preserves detail in both highlight and low-light areas but also avoids the tone mapping issues common in traditional HDR image synthesis, resulting in a more natural and vivid HDR image.

[0052] 6. Reduced hardware costs. Compared to specially designed multiple-exposure camera systems, the sCMOS sensor-based dual-gain HDR imaging system of the present invention utilizes the dual-gain output characteristics of existing sensors, eliminating the need for additional hardware modifications or complex mechanical shutter control, thereby reducing the system's hardware cost and complexity.

[0053] 7. Enhanced system robustness. Because the method of this invention relies on the inherent characteristics of the sCMOS sensor, rather than complex external mechanical structures or multiple exposure control, the robustness of the system in practical applications is significantly enhanced. This reduces the risk of mechanical failure and improves system reliability and stability.

[0054] In summary, the dual-gain HDR synthesis method and system based on the sCMOS sensor of the present invention realizes an efficient, accurate, and real-time HDR imaging solution by utilizing the dual-gain reading characteristics of the sensor, overcomes many defects of the traditional multiple exposure method, and has significant application value and technical advantages.

[0055] These and other aspects of the present invention will become more readily apparent in the following description of the embodiments. It should be understood that the above general description and the following detailed description are merely exemplary and explanatory and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the following briefly introduces the drawings required for the exemplary embodiments or related technical descriptions. The drawings are used to provide a further understanding of the present invention and constitute part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the drawings:

[0057] Figure 1 FIG. 4 is a flow chart of a dual-gain HDR synthesis method for an sCMOS sensor according to an embodiment of the present invention.

[0058] Figure 2 FIG. 4 is a schematic diagram of an ideal dual-gain channel relationship in a dual-gain HDR synthesis method for an sCMOS sensor according to an embodiment of the present invention.

[0059] Figure 3 This is a graph showing the relationship between the grayscale values ​​of high and low gain channels in the dual-gain HDR synthesis method of the sCMOS sensor according to an embodiment of the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence, nor do they limit "first" and "second" to be different types.

[0062] The following will clearly and completely describe the technical solutions in the exemplary embodiments of the present invention in conjunction with the accompanying drawings of the exemplary embodiments of the present invention. Obviously, the exemplary embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0063] While traditional multiple-exposure HDR synthesis methods present image registration challenges, the dual-gain characteristics of sCMOS sensors offer a new solution for HDR image synthesis. sCMOS sensors can simultaneously output two images, one high-gain and one low-gain. Through appropriate algorithmic synthesis, images with a wide dynamic range can be achieved. The present invention provides a dual-gain HDR synthesis method and system for sCMOS sensors. By performing a linear grayscale relationship analysis and grayscale stretching on high-gain and low-gain images, high-dynamic-range images can be rapidly synthesized. This approach aims to eliminate image registration issues and simplify the HDR image synthesis process. By performing a linear grayscale relationship analysis and grayscale stretching on high-gain and low-gain images, high-dynamic-range images can be rapidly synthesized.

[0064] The technical solution of the present invention is further described below with reference to specific embodiments:

[0065] See Figure 1 As shown, Figure 1 A flow chart of a dual-gain HDR synthesis method for an sCMOS sensor provided by the present invention. A dual-gain HDR synthesis method for an sCMOS sensor provided in one embodiment of the present invention includes the following steps:

[0066] Step S10: using the sCMOS sensor to simultaneously acquire two images, one with high gain and one with low gain.

[0067] Step S20: performing grayscale linear relationship analysis on the high-gain and low-gain images.

[0068] Step S30: Perform image histogram statistics to determine gain and exposure time settings.

[0069] Step S40: Perform grayscale stretching processing on the high-gain and low-gain images, and synthesize an HDR image based on the grayscale value relationship.

[0070] In this embodiment, the dual-gain HDR synthesis method and system for an sCMOS sensor uses an sCMOS sensor to simultaneously capture both high-gain and low-gain images. The high-gain image is used to capture details in darker areas of the scene, while the low-gain image is used to capture details in brighter areas. The sCMOS sensor internally amplifies the data through dual-gain channels, then quantizes the output through an ADC converter, creating an off-chip synthesis of images with a higher dynamic range.

[0071] In this embodiment, if the voltage amplitude corresponding to the full well charge is V m , then the output voltage range after the analog-to-digital converter is expressed as [0, V m ], that is, the signal voltage entering the column bus is V∈[0,V m ], define the gain channel magnification as G, KADC (V / DN) as the ADC converter unit, then the expression of the output data D after a bit precision quantization is:

[0072]

[0073] When V is at its minimum value, i.e., the system noise, in order to quantize V into a digital value for reading, at least At this time, the larger the value of gain G is, the smaller the signal data is read out.

[0074] The relationship between the output quantization data of the dual gain channel and the number of electrons in a single pixel after photoelectric conversion is shown in the figure below: Figure 2 As shown in the figure, the horizontal axis represents the number of electrons Q in a single pixel after photoelectric conversion, and the vertical axis represents the value D after quantization by ADC. For image information, it is used to represent the grayscale value of the corresponding pixel. H Indicates high gain channel, G L Indicates the low gain channel, F Q Represents the full well charge number. Figure 2 It can be seen from the figure that when the pixels are not saturated, the grayscale value relationship of the corresponding points in the high-gain image and the low-gain image is analyzed. In the non-saturated case, the grayscale value output by the two gain channels is approximately equal to the set gain value.

[0075] The dynamic range of the data read out for a single channel is 2 a -1:1. In the sCMOS sensor of this embodiment, when dual-channel image reading is adopted, the dynamic range of readable data is defined as:

[0076]

[0077] Where, max(D L ) represents the maximum quantization value of the low gain channel output, the maximum quantization value is 2 a -1, min(D H ) represents the minimum data that can be quantified by the high-gain channel. The minimum data that can be quantified by the high-gain channel is 0. The dynamic range of the image data read out by the sensor with dual-gain channel design is improved by 20*olog(G H / G L ), an HDR image with better visual effect is obtained by synthesizing the LDR images output by the two channels.

[0078] In this embodiment, when performing image histogram statistics and determining the gain and exposure time settings, if the grayscale range of abit is evenly divided into b grayscale levels, where a and b are both even numbers, then the number of dark pixels (grayscale value 0) is defined as S low ), saturated pixel (gray value is 2 a -1) (S high ), calculate the number of pixels in the upper and lower parts of the histogram of the two images as:

[0079]

[0080] Where P is the percentage of pixels in the histogram, s represents low gain, l represents high gain, H represents the image histogram, h represents the grayscale level (1-b), and N represents the number of all pixels. t,s ≈80%, P t,l ≈80%, then the gain and exposure time are adjusted.

[0081] Among them, when the pixels are not saturated, the grayscale value relationship between the corresponding points of the high-gain image and the low-gain image is:

[0082] H x，y =k*L x，y +b H x，y <2 a -1

[0083] Where H x,y Represents the grayscale value of high-gain image, L x,y Represents the grayscale value of the low-gain image, k and b represent coefficients. Using multiple sets of images for verification, the grayscale value relationship curve of the high and low gain channels is plotted as follows: Figure 3 As shown, the horizontal axis is the low-gain brightness value, and the vertical axis represents the high-gain brightness value. From the corresponding relationship between multiple groups of different gain combinations, it can be concluded that in the unsaturated area, the two show an obvious linear relationship, and this relationship is still valid for shooting different scenes, that is, each group of gain settings results in a fixed set of k and b.

[0084] In this embodiment, grayscale stretching is performed on high-gain and low-gain images. When synthesizing an HDR image based on the grayscale value relationship, the idea of ​​directly synthesizing the HDR image with useful information in the high- and low-gain images is adopted. In combination with the high- and low-gain grayscale characteristics, when the following formula is satisfied:

[0085]

[0086] Then {SH∪HL} can obtain an HDR image that retains the details of both images and the synthesized image is smooth. The algorithm proposed in this invention uses Lapray's method to obtain the original image, so it can directly set a H =S / 2, then the segmentation point a of the low-gain image can be directly obtained according to the formula L At the same time, in order to better display the image effect, grayscale stretching is added in the fusion process to achieve the effect of improving the image contrast.

[0087] Among them, linear stretching is used to x,y ≤a L and L x,y ≥a L The area is stretched so that the grayscale range of the fused image is distributed in the entire [0, S], then:

[0088]

[0089] Where H′ and L′ represent the images of H and L after a certain grayscale linear transformation respectively;

[0090] The synthesized image GHDR is represented as:

[0091]

[0092] In the dual-gain HDR synthesis method of the sCMOS sensor of the present invention, the dual-gain characteristics of the sCMOS sensor are utilized to simultaneously acquire high-gain and low-gain images at the same time, fundamentally avoiding the image registration problem caused by multiple exposures. Since the two images are acquired within the same frame time, there is no image inconsistency problem caused by the movement of objects in the scene, and ghosting and artifacts caused by registration errors are eliminated. By analyzing the grayscale value relationship between the high-gain and low-gain images and performing corresponding grayscale stretching and synthesis processing, the system can achieve high dynamic range imaging under single exposure. The high-gain image captures low-brightness details, while the low-gain image retains the high-brightness part details. The organic combination of the two can cover a wider brightness range, allowing the image to simultaneously display the dark and bright details in the scene.

[0093] Compared with traditional multiple-exposure methods, this system's approach does not require complex image registration and alignment processing, simplifying the HDR image synthesis process. Through precise grayscale value linear relationship analysis and grayscale stretching processing, the fusion of high-gain and low-gain images can be completed quickly and efficiently, reducing computational complexity and improving processing efficiency. Because sCMOS sensors can simultaneously output high-gain and low-gain images within a single frame, this method can support real-time HDR imaging. This is particularly important for applications in dynamic scenes, such as video surveillance, real-time scientific experimental imaging, and real-time medical imaging, and can significantly improve the system's response speed and imaging quality.

[0094] The present invention can effectively improve the contrast and detail expression of the image through grayscale stretching and synthesis strategies. The combination of high-gain images and low-gain images not only retains the details of the highlight and low-light areas, but also avoids the common tone mapping problems in traditional HDR image synthesis, making the final generated HDR image more visually natural and vivid. Compared with the multi-exposure camera system that requires a specially designed design, the dual-gain HDR imaging system based on the sCMOS sensor of the present invention utilizes the dual-gain output characteristics of the existing sensor, without the need for additional hardware modification or complex mechanical shutter control, thereby reducing the hardware cost and complexity of the system. Since the method of the present invention relies on the characteristics of the sCMOS sensor itself, rather than an external complex mechanical structure or multi-exposure control, the robustness of the system in practical applications is significantly enhanced. The risk of mechanical failure is reduced and the reliability and stability of the system are improved.

[0095] It should be understood that, although the above is described in a certain order, these steps are not necessarily performed in sequence according to the above order. Unless there is clear explanation in this article, the execution of these steps does not have strict order restriction, and these steps can be performed in other orders. Moreover, a part of the steps of the present embodiment may include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be performed in turn or alternately with at least a portion of the steps or stages in other steps or other steps.

[0096] In one embodiment, the present invention provides a dual-gain HDR synthesis system for an sCMOS sensor, configured to execute the dual-gain HDR synthesis method for the sCMOS sensor. The dual-gain HDR synthesis method includes:

[0097] Dual-gain image acquisition module, used to simultaneously acquire high-gain and low-gain images using an sCMOS sensor;

[0098] Gray value linear relationship analysis module, used to perform gray value linear relationship analysis on high gain and low gain images;

[0099] Image histogram statistics module, used to perform image histogram statistics and determine gain and exposure time settings;

[0100] Grayscale stretching module, used to perform grayscale stretching processing on high-gain and low-gain images;

[0101] The HDR image synthesis module is used to synthesize HDR images based on the grayscale value relationship.

[0102] In this embodiment, the camera adopts the steps of the aforementioned dual-gain HDR synthesis method for an sCMOS sensor during execution. Therefore, the operation process of the camera will not be described in detail in this embodiment.

[0103] In summary, the dual-gain HDR synthesis method and system for sCMOS sensors of the present invention, by utilizing the dual-gain reading characteristics of the sensor, realizes an efficient, accurate, and real-time HDR imaging solution, overcomes many shortcomings of traditional multiple exposure methods, and has significant application value and technical advantages.

[0104] In one embodiment, the present invention further provides a computer device comprising at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the at least one processor to perform the steps of the dual-gain HDR synthesis method for the sCMOS sensor.

[0105] In one embodiment, the present invention further provides a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the steps of the dual-gain HDR synthesis method for the sCMOS sensor.

[0106] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program represented by computer instructions. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above-mentioned methods. In addition, any reference to memory, storage, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory.

[0107] Non-volatile memory can include read-only memory, magnetic tape, floppy disk, flash memory, or optical storage. Volatile memory can include random access memory or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory or dynamic random access memory.

[0108] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A dual-gain HDR synthesis method for an sCMOS sensor, characterized in that: The method comprises the following steps: Use sCMOS sensor to acquire high-gain and low-gain images simultaneously; Perform linear relationship analysis on grayscale values ​​of high-gain and low-gain images; Perform image histogram statistics to determine gain and exposure time settings; Perform grayscale stretching on high-gain and low-gain images, and synthesize HDR images based on the grayscale value relationship; Among them, the useful information in the high and low gain images is directly synthesized into HDR images, and the high and low gain grayscale characteristics are combined when the following formula is satisfied: ， Use linear stretching to as well as Stretch the area.

2. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 1, wherein: When using an sCMOS sensor to simultaneously acquire high-gain and low-gain images, the high-gain image is used to capture details in darker parts of the scene, and the low-gain image is used to capture details in brighter parts of the scene.

3. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 2, wherein: The sCMOS sensor is internally amplified by dual gain channels and then quantized by an ADC converter to produce an image with a higher dynamic range off-chip.

4. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 1, wherein: In sCMOS sensors, if the voltage amplitude corresponding to the full well charge is , then the output voltage range after the analog-to-digital converter is expressed as [0, ], that is, the signal voltage entering the column bus is V ∈ [0, ], define the gain channel amplification factor as , If ADC is the unit of the converter, the expression of the output data D after a bit precision quantization is: ， when When the value is the minimum, that is, the system noise, Quantization into digital reading must at least meet , at this time the gain The larger the value, the smaller the signal data will be read.

5. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 4, wherein: When the pixels are not saturated, the grayscale value relationship between the corresponding points of the high-gain image and the low-gain image is analyzed. When the pixels are not saturated, the grayscale values ​​output by the two gain channels are approximately equal to the set gain values.

6. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 5, wherein: In an sCMOS sensor, when dual-channel image reading is used, the dynamic range of readable data is defined as: ， Where, Indicates the maximum quantized value of the low gain channel output. The maximum quantized value is , Indicates the minimum data that can be quantified by the high-gain channel. The minimum data that can be quantified by the high-gain channel is 0. The dynamic range of the image data read by the sensor with dual-gain channel design is improved compared with the single-channel design. , the HDR image is obtained by synthesizing the LDR images output by the two channels.

7. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 6, wherein: When performing image histogram statistics and determining the gain and exposure time settings, if the grayscale range of a bit is evenly divided into b grayscale levels, where a and b are both even numbers, then the number of dark pixels is defined as , the gray value is 0, and the gray value of the saturated pixel is Number of , calculate the number of pixels in the upper and lower parts of the histogram of the two images as: ， ， Where P is the percentage of pixels in the histogram, s represents low gain, l represents high gain, H represents the image histogram, h represents the grayscale level (1-b), and N represents the number of all pixels. 、 , then the gain and exposure time are adjusted; Indicates the percentage of pixels in the low-gain histogram; Represents the number of pixels with grayscale level (1-b) in the low-gain histogram; Indicates the percentage of pixels in the high gain histogram; Indicates the number of pixels with grayscale level (1-b) in the high gain histogram.

8. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 7, wherein: When the pixels are not saturated, the grayscale value relationship between the corresponding points of the high-gain image and the low-gain image is: ， Where, Represents the grayscale value of the high-gain image, Represents the grayscale value of the low-gain image, k and b represent coefficients, and each set of gain settings results in a fixed set of k and b.

9. The dual-gain HDR synthesis method for an sCMOS sensor according to claim 8, wherein: Perform grayscale stretching on the high-gain and low-gain images. When synthesizing the HDR image based on the grayscale value relationship, the grayscale range of the fused image is distributed in the entire [0, S]. Then: ， ， Where, 、 Represented as H, L respectively after grayscale linear transformation images; The synthesized image GHDR is represented as: ， 。 10. A dual-gain HDR synthesis system for an sCMOS sensor, characterized in that: A dual-gain HDR synthesis method for an sCMOS sensor according to any one of claims 1 to 9, the system comprising: Dual-gain image acquisition module, used to simultaneously acquire high-gain and low-gain images using an sCMOS sensor; Gray value linear relationship analysis module, used to perform gray value linear relationship analysis on high gain and low gain images; Image histogram statistics module, used to perform image histogram statistics and determine gain and exposure time settings; Grayscale stretching module, used to perform grayscale stretching processing on high-gain and low-gain images; The HDR image synthesis module is used to synthesize the HDR image according to the grayscale value relationship. The useful information in the high and low gain images is directly synthesized into the HDR image, and the high and low gain grayscale characteristics are combined to meet the following formula: ， Use linear stretching to as well as Stretch the area.

Images