Image processing method and system for parameter adjustment based on feedback
By adjusting the parameters of the image signal processor through bit-width mapping and feedback, the problem that the image processing flow in the prior art cannot adapt to different perception models is solved, and the perception model can effectively identify image targets.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- MOMENTA (SUZHOU) TECHNOLOGY CO LTD
- Filing Date
- 2021-12-30
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technologies cannot select image processing procedures that are suitable for different perception models, resulting in unsatisfactory target representation in images and affecting the recognition performance of perception models.
The high-bit-width image is mapped to a medium-bit-width image through bit-width mapping logic, and then processed into a low-bit-width RGB image by the image signal processor. After the perception model identifies the region of interest, it generates feedback information and adjusts the parameters of the image signal processor so that the processed image meets the requirements of the perception model.
This improves the recognition performance of the perception model for targets in images, meets the needs of different perception models, and ensures that the characteristics of the region of interest meet the requirements.
Smart Images
Figure CN116416144B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of image processing technology, and in particular to an image processing method and system based on feedback for parameter adjustment. Background Technology
[0002] Because the dynamic range of the images output by the image sensors in the onboard cameras of autonomous vehicles is relatively high, the ISP (Image Signal Processor) will compress part of the dynamic range. This inevitably results in a loss of dynamic range, which is a loss of contrast.
[0003] To address the dynamic range loss issue, different perceptual models may focus on different aspects. Some models might focus on brighter targets in the image, while others might focus on darker targets. Even after all images are processed by the same image signal processor, it's impossible to achieve the ideal image state for each perceptual model's specific target. Furthermore, if the target itself in the image doesn't appear ideal, it will also affect the perceptual model's recognition performance. Summary of the Invention
[0004] To address the problem that existing technologies cannot select appropriate image processing procedures to adapt to different needs of perception models, this application mainly provides an image processing method and system based on feedback-based parameter adjustment.
[0005] One technical solution adopted in this application is: providing an image processing method for parameter adjustment based on feedback, which includes:
[0006] The bit-width mapping logic maps the high bit-width image acquired by the image sensor to the medium bit-width image required by the image signal processor.
[0007] The image signal processor processes the medium-width image and converts it into a low-width RGB image.
[0008] The perception model identifies low-bit-width RGB images, determines the region of interest, and generates image signal processor parameter adjustment feedback information containing relevant information about the region of interest to the image signal processor.
[0009] The image signal processor adjusts its parameters based on feedback information, ensuring that the region of interest obtained after processing the median-width image has the characteristics required by the perceptual model.
[0010] Another technical solution adopted in this application is: providing an image processing system based on feedback parameter adjustment, which includes bit-width mapping logic, an image signal processor, and a perceptual model, wherein:
[0011] The bit-width mapping logic maps the high bit-width image acquired by the image sensor to the medium bit-width image required by the image signal processor.
[0012] The image signal processor processes the medium-width image and converts it into a low-width RGB image.
[0013] The perception model identifies low-bit-width RGB images, determines the region of interest, and generates image signal processor parameter adjustment feedback information containing relevant information about the region of interest to the image signal processor.
[0014] The image signal processor adjusts its parameters based on feedback information, ensuring that the region of interest obtained after processing the median-width image has the characteristics required by the perceptual model.
[0015] Another technical solution adopted in this application is to provide a computer-readable storage medium storing computer instructions that are operated to perform the image processing method based on feedback parameter adjustment in Solution 1.
[0016] Another technical solution adopted in this application is: providing a computer device, which includes a processor and a memory, the memory storing computer instructions, which are operated to execute the image processing method based on feedback parameter adjustment in Solution 1.
[0017] The beneficial effects achievable by the technical solution of this application are as follows: This application designs an image processing method and system based on feedback-driven parameter adjustment. The perception model in this application identifies targets in images processed by an image signal processor (ISP). Based on the perception effect of the perception model, the perception model reverse-configures the corresponding ISP parameters according to the processing flow of the ISP. By processing the image using the configured ISP parameters, the perception model can better identify targets in the processed image. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of this application 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 some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0019] Figure 1 This is a schematic diagram of a specific implementation of an image processing method for parameter adjustment based on feedback according to this application;
[0020] Figure 2 This is a schematic diagram of a specific implementation of an image processing system based on feedback parameter adjustment according to this application.
[0021] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation
[0022] The preferred embodiments of this application will now be described in detail with reference to the accompanying drawings, so that the advantages and features of this application can be more easily understood by those skilled in the art, thereby providing a clearer and more definite definition of the scope of protection of this application.
[0023] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.
[0024] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.
[0025] Figure 1 This illustration shows a specific implementation of an image processing method based on feedback parameter adjustment according to this application. Figure 1 In the specific implementation shown, the image processing method based on feedback for parameter adjustment mainly includes:
[0026] Step S101: The high bit-width image acquired by the image sensor is mapped to the medium bit-width image required by the image signal processor by the bit-width mapping logic.
[0027] In this embodiment, the original image output from the image sensor in the vehicle camera has a high bit width dynamic range, which is beneficial for algorithm recognition. However, the large bit width results in an excessive amount of data that cannot be processed. Therefore, it is necessary to compress the dynamic range of the bit width of the original image with a high bit width dynamic range according to the bit width mapping logic.
[0028] In a specific example of this application, the image signal processor (Image Signal Processor) prunes the bit width of a high-bit-width dynamic image based on the ISP's allowed bit width; or it maps the bit width of the high-bit-width dynamic image based on a mapping function provided by the Image Signal Processor. When the bit width of the high-bit-width dynamic image is greater than the ISP's allowed bit width, the bit width needs to be compressed, and the compression methods include bit pruning and function mapping. The image signal processor's back-end algorithm directly prunes the bit width of the high-bit-width dynamic image, either pruning the lower bits or the higher bits. Based on the mapping function fed back by the Image Signal Processor, details in the dark areas of the image can be preserved, and the overall image can be prevented from being overexposed.
[0029] exist Figure 1 In the specific implementation shown, the image processing method for adjusting parameters based on feedback further includes:
[0030] In step S102, the image signal processor processes the mid-width image and converts it into a low-width RGB image.
[0031] In this embodiment, the image signal processor processes the mid-width image using various parameters to transform it into a low-width RGB image, enabling the perception model to better perceive the target in the image.
[0032] In one optional embodiment of this application, the median-width image is processed by an image signal processor, including: performing white balance processing, tone mapping processing, demosaic processing, and gamma correction processing on the median-width image in sequence, wherein the parameters of the image signal processor include white balance parameters, tone mapping parameters, demosaic parameters, and / or gamma correction parameters.
[0033] In this embodiment, the parameter processing of the image signal processor (ISP) includes, but is not limited to, blacklevel compensation, lens shading correction, bad pixel correction, color interpolation, Bayer noise removal, white balance (AWB) correction, color correction, and gamma correction. This results in better image quality.
[0034] In a specific example of this application, the image signal processor (ISP) performs white balance on the median-width image by adjusting the ratio of RGB pixels to restore the white pixel values in the median-width image to white, thus obtaining a white-balanced median-width image; the image signal processor (ISP) performs tone mapping on the white-balanced median-width image to obtain a grayscale image; the image signal processor (ISP) performs de-mosaicing on the grayscale image to obtain a color image; the image signal processor (ISP) performs gamma correction on the color image, and finally outputs a low-bit-width RGB image.
[0035] In a specific example of this application, white balance (AWB) correction is crucial. If an image hasn't been white-balanced, the white might not actually be white. The human eye perceives different proportions of red, green, and blue in reflected light depending on the scene and lighting. The white balance function in the ISP module essentially performs a similar function to the human eye. If the white of the scene being adjusted leans slightly towards blue, the white balance adjustment changes the preset proportions to ensure the resulting image remains white. The biggest advantage of automatic white balance is its simplicity and speed. Tone mapping, because image sensors output photoelectric signals, has a linear mapping relationship: the strength of the light signal directly correlates to the pixel value of the converted electrical signal. However, for the human eye, it's not a linear mapping; it's a non-linear one, and the non-linear mapping varies depending on the target. Therefore, tone mapping also simulates the function of the human eye to perform a non-linear mapping. Demosaicing occurs because the pixels output by an image sensor are RGB (radix, RGB), and each pixel needs to be converted to have RGB information. Gamma correction, while essentially linear data, appears non-linear to the human eye. This involves remapping the entire image using a non-linear method, similar to applying a gamma curve or an inverse logarithmic curve.
[0036] In an optional embodiment of this application, the image signal processor calculates the brightness level of the mid-bit-width image in blocks, including: dividing the high-bit-width image into a preset number of blocks, calculating the weighted average brightness of the pixels in each block, and obtaining the brightness level of each block.
[0037] In this embodiment, the image signal processor divides the median-width image into a preset number of blocks, calculates and obtains the brightness value of each block, i.e., the weighted average of the brightness of the pixels, and then obtains the brightness level of the median-width image.
[0038] exist Figure 1In the specific implementation shown, the image processing method for adjusting parameters based on feedback further includes:
[0039] Step S103: The perception model identifies the low-bit-width RGB image, determines the region of interest, and generates image signal processor parameter adjustment feedback information containing relevant information about the region of interest to the image signal processor.
[0040] In this embodiment, the perception model detects targets in the input low-bit-width RGB image and obtains feedback information for adjusting the parameters of the image signal processor. The perception model can perceive targets that it needs to focus on. The perception model perceives the low-bit-width RGB image processed by the image signal processor, and obtains some image signal processor parameters based on the strength of the perception effect to adjust the feedback information.
[0041] In one optional embodiment of this application, an image signal processor calculates the brightness level of the mid-width image in blocks, and determines the required grayscale accuracy for the brightness of different regions of the mid-width image based on the region of interest fed back by the perception model and the corresponding brightness level of the region of interest in the mid-width image. Based on the required grayscale accuracy for the brightness of different regions of the mid-width image, it generates bit-width mapping feedback information for the bit-width mapping logic. Based on the bit-width mapping feedback information, the bit-width mapping logic adjusts the mapping relationship between the high-width image and the mid-width image to ensure that the brightness of the region of interest has high grayscale accuracy.
[0042] In this embodiment, the bit-width mapping feedback information includes the grayscale precision of each patch in the mid-bit-width image. Based on the brightness level of the region of interest (ROI) corresponding to the mid-bit-width image in the image signal processor (ESP) according to the feedback from the perceptual model, the grayscale precision is determined and bit-width mapping feedback information is generated. The ESP then generates a mapping function based on this information, which preserves details in the dark areas of the original image while ensuring the overall image is not overexposed. The mapping relationship between the high-bit-width and mid-bit-width images is adjusted based on the bit-width mapping feedback information, making it easier for the perceptual model to identify the ROI.
[0043] In one optional embodiment of this application, the bit-width mapping logic adjusts the mapping relationship between the high-bit-width image and the mid-bit-width image based on the bit-width mapping feedback information. This includes: when the brightness of the region of interest is low, the bit-width mapping feedback information instructs the bit-width mapping logic to cut off one or more consecutive bits starting from the highest bit of the high-bit-width image to obtain the mid-bit-width image; or when the brightness of the region of interest is high, the bit-width mapping feedback information instructs the bit-width mapping logic to cut off one or more consecutive bits starting from the lowest bit of the high-bit-width image to obtain the mid-bit-width image.
[0044] In this embodiment, based on the bit-width mapping feedback information, a medium bit-width image is obtained by directly cutting bits from a high bit-width image. Bit cutting can only cut low bits or high bits. Bit cutting is used to perform bit-width mapping to preserve the details of targets with relatively low pixel values in the high bit-width image as much as possible.
[0045] In an optional embodiment of this application, the bit-width mapping logic adjusts the mapping relationship between the high bit-width image and the medium bit-width image based on the bit-width mapping feedback information. This includes: when the brightness of the region of interest contains multiple brightness levels, the bit-width mapping feedback information instructs the bit-width mapping logic to obtain the medium bit-width image using all bits of the high bit-width image according to the bit-width mapping function specified by the image signal processor.
[0046] In this embodiment, according to the mapping relationship in the bit-width mapping function specified by the image signal processor, the high bit-width image is used as the input of the bit-width mapping function to obtain the output of the bit-width mapping function, namely the medium bit-width image.
[0047] In a specific example of this application, the perceptual model detects targets in a low-bit-width RGB image and filters out targets of interest (ROIs) within the image. Based on the ROIs, the range of the ROIs in the low-bit-width RGB image is obtained, ultimately yielding perceptual image signal processor parameter adjustment feedback information for the range of the ROIs contained in the low-bit-width RGB image. The perceptual model can determine the image delineation region based on the image processed by the image signal processor and modify the image signal processor parameters based on the delineated values in the image.
[0048] exist Figure 1 In the specific implementation shown, the image processing method for adjusting parameters based on feedback further includes:
[0049] In step S104, the image signal processor adjusts its parameters according to the feedback information, so that the region of interest obtained after the median width image is processed by the image signal processor has the characteristics required by the perception model.
[0050] In this embodiment, the perception model feeds back the parameter adjustment feedback information of the image signal processor to the image signal processor. The image signal processor then adjusts its parameters accordingly, resulting in better image processing and fulfilling the characteristics required by the perception model.
[0051] In one optional embodiment of this application, the image signal processor adjusts the parameters of the image signal processor according to the image signal processor parameter adjustment feedback information. This includes adjusting at least one of the white balance parameters, tone mapping parameters, demosaic parameters, and / or gamma correction parameters according to the image signal processor parameter adjustment feedback information, so that the region of interest obtained after the median width image is processed by the image signal processor has the characteristics required by the perception model.
[0052] In this embodiment, feedback information is adjusted based on the image signal processor parameters generated according to the region of interest, adjusting at least one parameter in the image signal processor. The specific adjustment is made based on the actual feedback results.
[0053] In a specific embodiment of this application, the image signal processor (ISP) uses the ISP algorithm to identify the feedback information of the perceived image signal processor parameter adjustment, obtaining the identified adjustment information. Based on the identified adjustment information, the parameters of the image signal processor are adjusted accordingly. The ISP algorithm at the back end of the image signal processor can identify the image signal processor parameter adjustment feedback information from the perception model and modify the image signal processor parameters accordingly. The image signal processor processes the next frame of high-bit-width dynamic image based on the adjusted image signal processor parameters and outputs the processing result to the perception model. The adjusted image signal processor parameters result in better processing of the next frame of high-bit-width dynamic image, enabling the perception model to better perceive targets in the image.
[0054] In a specific example of this application, a corresponding image processing flow is selected based on the different image requirements of different types of perception models. This image processing flow includes various pipelines within an image signal processor (Image Signal Processor) and a processing flow without an Image Signal Processor. By modifying the parameters of the Image Signal Processor corresponding to different pipelines within the Image Signal Processor using a feedback mechanism, the images processed by the Image Signal Processor can better meet the requirements of different types of perception.
[0055] Figure 2 This illustration shows a specific implementation of an image processing system based on feedback parameter adjustment according to this application. Figure 2 In the specific implementation shown, the image processing system that adjusts parameters based on feedback mainly includes module 201 (bit-width mapping logic), module 202 (image signal processor), and module 203 (perceptual model), wherein:
[0056] The high-bit-width image acquired by the image sensor is mapped to the medium-bit-width image required by the image signal processor by the bit-width mapping logic of module 201.
[0057] The image signal processor in module 202 processes the medium-width image and converts it into a low-width RGB image.
[0058] The perceptual model of module 203 identifies the low bit-width RGB image, determines the region of interest, and generates image signal processor parameter adjustment feedback information containing relevant information about the region of interest to the image signal processor.
[0059] The image signal processor module 202 adjusts the parameters of the image signal processor based on the parameter adjustment feedback information, so that the region of interest obtained after the median width image is processed by the image signal processor has the characteristics required by the perception model.
[0060] In this embodiment, the perception model senses the low-bit-width RGB image output by the image signal processor, and feeds back the parameters of different image signal processors. The image signal processor adjusts its own parameters and performs the same processing on the next frame image, making the image clearer and the perception model more likely to identify its own target of interest.
[0061] In an optional embodiment of this application, the image signal processor calculates the brightness level of the mid-bit-width image in blocks, including: dividing the high-bit-width image into a preset number of blocks, calculating the weighted average brightness of the pixels in each block, and obtaining the brightness level of each block.
[0062] In this embodiment, the image signal processor divides the median-width image into a preset number of blocks, calculates and obtains the brightness value of each block, i.e., the weighted average of the brightness of the pixels, and then obtains the brightness level of the median-width image.
[0063] In one optional embodiment of this application, the median-width image is processed by an image signal processor, including: performing white balance processing, tone mapping processing, demosaic processing, and gamma correction processing on the median-width image in sequence, wherein the parameters of the image signal processor include white balance parameters, tone mapping parameters, demosaic parameters, and / or gamma correction parameters.
[0064] In this embodiment, the parameter processing of the image signal processor (ISP) includes, but is not limited to, blacklevel compensation, lens shading correction, bad pixel correction, demosaic, Bayer noise removal, white balance (AWB) correction, color correction, and gamma correction. This results in better image quality.
[0065] In one optional embodiment of this application, the image signal processor adjusts the parameters of the image signal processor according to the image signal processor parameter adjustment feedback information. This includes adjusting at least one of the white balance parameters, tone mapping parameters, demosaic parameters, and / or gamma correction parameters according to the image signal processor parameter adjustment feedback information, so that the region of interest obtained after the median width image is processed by the image signal processor has the characteristics required by the perception model.
[0066] In this embodiment, feedback information is adjusted based on the image signal processor parameters generated according to the region of interest, adjusting at least one parameter in the image signal processor. The specific adjustment is made based on the actual feedback results.
[0067] The image processing system for parameter adjustment based on feedback provided in this application can be used to execute the image processing method for parameter adjustment based on feedback described in any of the above embodiments. Its implementation principle and technical effect are similar, and will not be repeated here.
[0068] In one specific embodiment of this application, the functional modules of the image processing system based on feedback for parameter adjustment can be directly in hardware, in software modules executed by a processor, or in a combination of both.
[0069] Software modules may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in this art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium.
[0070] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination thereof. A general-purpose processor can be a microprocessor, but alternatively, it can be any conventional processor, controller, microcontroller, or state machine. The processor can also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors incorporating a DSP core, or any other such configuration. Alternatively, the storage medium can be integrated with the processor. The processor and storage medium can reside in an ASIC. The ASIC can reside in the user terminal. Alternatively, the processor and storage medium can reside as discrete components in the user terminal.
[0071] In another specific embodiment of this application, a computer-readable storage medium stores computer instructions that are operated to perform the image processing method for parameter adjustment based on feedback in any embodiment.
[0072] In another specific embodiment of this application, a computer device includes a processor and a memory, the memory storing computer instructions that are operated to perform the image processing method based on feedback parameter adjustment in any embodiment.
[0073] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.
[0074] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0075] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. An image processing method based on feedback for parameter adjustment, characterized in that, include: The bit-width mapping logic maps the high bit-width image acquired by the image sensor to the medium bit-width image required by the image signal processor. The image signal processor processes the median-width image and converts it into a low-bit-width RGB image. The processing of the median-width image by the image signal processor includes: The median-width image is sequentially processed with white balance, tone mapping, demosaicing, and gamma correction. The parameters of the image signal processor include white balance parameters, tone mapping parameters, demosaic parameters, and / or gamma correction parameters. The perception model identifies the low-bit-width RGB image, determines the region of interest, and generates image signal processor parameter adjustment feedback information containing information related to the region of interest to the image signal processor. The image signal processor adjusts its parameters based on parameter adjustment feedback information, so that the region of interest obtained after processing the median-width image by the image signal processor has the characteristics required by the perceptual model. The adjustment of the image signal processor parameters based on parameter adjustment feedback information includes: Based on the image signal processor parameter adjustment feedback information, at least one of the white balance parameter, tone mapping parameter, demosaic parameter, and / or gamma correction parameter is adjusted so that the region of interest obtained after the median width image is processed by the image signal processor has the characteristics required by the perception model.
2. The image processing method for parameter adjustment based on feedback as described in claim 1, characterized in that, Also includes: The image signal processor, The brightness level of the median-width image is calculated in blocks, and based on the region of interest fed back by the perception model and the corresponding brightness level of the region of interest in the median-width image, the required grayscale accuracy for the brightness of different regions of the median-width image is determined. Based on the grayscale accuracy required for the brightness of different regions of the mid-width image, bit-width mapping feedback information is generated for the bit-width mapping logic feedback. as well as The bit-width mapping logic adjusts the mapping relationship between the high bit-width image and the medium bit-width image based on the bit-width mapping feedback information, so that the brightness of the region of interest has high grayscale accuracy.
3. The image processing method for parameter adjustment based on feedback as described in claim 2, characterized in that, The step of adjusting the mapping relationship between the high-bit-width image and the mid-bit-width image by the bit-width mapping logic based on the bit-width mapping feedback information includes: When the brightness of the regions of interest is all at low levels, the bit-width mapping feedback information instructs the bit-width mapping logic to remove one or more consecutive bits starting from the highest bit of the high-bit-width image to obtain the medium-bit-width image; or When the brightness of the regions of interest is all at a high brightness level, the bit width mapping feedback information instructs the bit width mapping logic to cut off one or more consecutive bits starting from the least significant bit of the high bit width image to obtain the medium bit width image.
4. The image processing method for parameter adjustment based on feedback as described in claim 2, characterized in that, The step of adjusting the mapping relationship between the high-bit-width image and the mid-bit-width image by the bit-width mapping logic based on the bit-width mapping feedback information includes: When the brightness of the region of interest contains multiple brightness levels, the bit-width mapping feedback information instructs the bit-width mapping logic to obtain the medium bit-width image using all bits of the high bit-width image, according to the bit-width mapping function specified by the image signal processor.
5. The image processing method for parameter adjustment based on feedback as described in claim 2, characterized in that, The step of calculating the brightness level of the median-width image in blocks by the image signal processor includes: The high bit-width image is divided into a preset number of blocks, and the weighted average brightness of the pixels in each block is calculated to obtain the brightness level of each block.
6. An image processing system for parameter adjustment based on feedback, comprising bit-width mapping logic, an image signal processor, and a perceptual model, characterized in that, The bit-width mapping logic maps the high bit-width image acquired by the image sensor to the medium bit-width image required by the image signal processor. The image signal processor processes the median-width image and transforms it into a low-bit-width RGB image. The processing of the median-width image by the image signal processor includes sequentially performing white balance processing, tone mapping processing, demosaic processing, and gamma correction processing on the median-width image. The parameters of the image signal processor include white balance parameters, tone mapping parameters, demosaic parameters, and / or gamma correction parameters. The perception model identifies the low-bit-width RGB image, determines the region of interest, and generates image signal processor parameter adjustment feedback information containing information related to the region of interest to the image signal processor. The image signal processor adjusts its parameters according to the image signal processor parameter adjustment feedback information, so that the region of interest obtained after the median-width image is processed by the image signal processor has the characteristics required by the perception model. The adjustment of the image signal processor parameters by the image signal processor according to the image signal processor parameter adjustment feedback information includes: adjusting at least one of the white balance parameter, tone mapping parameter, demosaic parameter, and / or gamma correction parameter according to the image signal processor parameter adjustment feedback information, so that the region of interest obtained after the median-width image is processed by the image signal processor has the characteristics required by the perception model.
7. A computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are operated to perform the image processing method based on feedback parameter adjustment as described in any one of claims 1-5.
8. A computer device comprising a processor and a memory storing computer instructions, wherein the processor operates the computer instructions to perform the image processing method for feedback-based parameter adjustment as described in any one of claims 1-5.