Camera module automated debugging system with adaptive parameter calibration

Abstract

The present invention discloses an automated debugging system for camera modules with adaptive parameter calibration, which relates to the field of camera debugging technology. The parameter calibration module sets physical parameter thresholds based on the specifications of the camera module, and then uses a calibration plate to calibrate the geometric and optical parameters to generate the boundary conditions of the camera module. The adaptive debugging module detects the environmental parameters of the shooting environment in real time and dynamically adjusts the imaging parameters of each camera. The multi-camera coordination module coordinates exposure across cameras and uses global white balance mapping to align curves according to the color temperature response curves of multiple cameras. Combined with a machine learning algorithm, the color data of multiple cameras is matched with color styles, and the camera parameters are adjusted when multiple cameras are switched. The debugging system adopts a layered debugging architecture combined with environmental adaptive optimization and cross-camera parameter coordination to achieve an automated, multi-camera collaborative high-precision parameter calibration process, thereby improving the consistency and stability of the camera module.

Description

Technical Field

[0001] The present invention relates to the technical field of camera debugging, and in particular to an automatic debugging system for a camera module with adaptive parameter calibration. Background Art

[0002] With the development of smartphones, smart car systems, and security monitoring equipment, multi-camera modules (such as main camera, ultra-wide-angle camera, and telephoto camera) have become an important component of imaging systems. However, due to differences in optical characteristics, sensor characteristics, and imaging algorithms among different cameras, it is difficult to ensure imaging consistency in complex lighting conditions (such as backlight, night scenes, and dynamic scenes). The main manifestations are:

[0003] 1. Inconsistent white balance: The main camera and telephoto camera perceive different color temperatures in the same scene, resulting in color shift.

[0004] 2. Exposure parameter conflict: In a high dynamic range (HDR) environment, the exposure values ​​of different cameras deviate, resulting in uneven brightness.

[0005] 3. Differences in optical distortion: Ultra-wide-angle cameras have significant perspective distortion, while telephoto cameras focus more on long-range resolution, making it difficult to achieve uniform geometric correction.

[0006] 4. Insufficient environmental adaptability: Traditional debugging relies on static parameter tables, which makes it difficult to dynamically respond to different shooting scenarios, resulting in reduced image quality.

[0007] In order to solve the above problems, this application proposes an automatic debugging system for camera modules with adaptive parameter calibration, which adopts a layered debugging architecture combined with environmental adaptive optimization and cross-camera parameter coordination to realize an automated, multi-camera collaborative high-precision parameter calibration process, thereby improving the consistency and stability of the camera module. Summary of the Invention

[0008] The purpose of the present invention is to provide an automatic debugging system for a camera module with adaptive parameter calibration to solve the shortcomings of the background technology.

[0009] In order to achieve the above object, the present invention provides the following technical solution: an automatic debugging system for a camera module with adaptive parameter calibration, comprising a parameter calibration module, an adaptive debugging module, and a multi-camera coordination module;

[0010] Parameter calibration module: During the factory commissioning phase, physical parameter thresholds are set based on the camera module's specifications. A calibration plate is then used to calibrate the geometric and optical parameters to generate the camera module's boundary conditions.

[0011] Adaptive debugging module: real-time detection of environmental parameters of the shooting environment and dynamic adjustment of imaging parameters of each camera;

[0012] Multi-camera coordination module: This module coordinates exposure across cameras, uses global white balance mapping, aligns the color temperature response curves of multiple cameras, and combines machine learning algorithms to perform color style matching on color data from multiple cameras.

[0013] Preferably, the multi-camera coordination module performs exposure coordination across cameras and performs exposure parameter conversion, and the expression is: , where is the exposure value after conversion of the target camera, is the exposure value of the reference camera, is the aperture diameter of the reference camera, is the aperture diameter of the target camera, is the ISO sensitivity of the target camera, is the ISO sensitivity of the reference camera, is the exposure time of the target camera, is the exposure time of the reference camera.

[0014] Preferably, the multi-camera coordination module performs white balance gain calculation, and the expression is: , where is the white balance gain of the target camera, is the white balance gain of the reference camera, is the color temperature detected by the target camera, The color temperature detected by the reference camera.

[0015] Preferably, the multi-camera coordination module uses a machine learning model to analyze the color shift of different cameras under the same lighting conditions, and performs style unification based on color distribution histogram matching. The color histogram matching algorithm is: , where is the color adjustment value of the target camera, is the color value of the reference camera, 、 is the color histogram feature of the target camera and the reference camera.

[0016] Preferably, the adaptive debugging module adaptively adjusts the camera's imaging parameters and exposure time based on the environmental parameter detection results: , where is the adjusted exposure time, is the exposure time before adjustment, is the exposure brightness, is the adjustment coefficient;

[0017] Calculate the new RGB gains: , where is the adjusted RGN gain, is the RGN gain before adjustment, is the standard color temperature, is the currently detected color temperature;

[0018] Calculating ISO settings: , For the adjusted value, Before adjustment value, is the target brightness, is the exposure brightness.

[0019] Preferably, the adaptive debugging module detects the scene brightness through the exposure value of the RGB sensor or the CMOS sensor and calculates the exposure brightness. The expression is: , where is the exposure brightness, is the average brightness value of the image, is the lens aperture area;

[0020] The gray world algorithm is used to calculate the scene color temperature, and the expression is: , where is the estimated color temperature, 、 、 is the average value of the three channels of the image;

[0021] Calculate the speed of a moving object using the expression: , where is the object speed, is the displacement of the object in two consecutive frames, is the frame interval time.

[0022] Preferably, the parameter calibration module sets thresholds of physical parameters according to the optical specifications and hardware characteristics of the camera module, including aperture size, focal length range, distortion compensation limit and photosensitivity;

[0023] Use a calibration plate to perform geometric parameter calibration, including focal length calibration and distortion correction.

[0024] Preferably, the parameter calibration module shoots a calibration plate grid pattern with an ultra-wide-angle camera, analyzes the degree of deformation of Barrel-Distortion and Pincushion-Distortion, uses an optical distortion correction algorithm, calculates compensation parameters, and optimizes the lens correction curve;

[0025] Under a standard color temperature light source, photograph a white reference plate, measure the response values ​​of the RGB channels, calculate the white balance gain parameters, and correct the white balance deviation of different cameras;

[0026] Use standard color cards for color comparison, analyze the camera's color reproduction capabilities under different light sources, calculate color difference, and adjust the color mapping matrix.

[0027] Preferably, the parameter calibration module tests the ISO sensitivity range of the camera in a low-light environment and sets the lowest available ISO for different cameras;

[0028] Use high-contrast test scenes to measure the dynamic range of the camera and set the HDR trigger threshold so that the camera automatically turns on HDR in high dynamic range scenes;

[0029] Generate the boundary conditions of the camera module based on the test results, including optical limit parameters, color correction matrix, sensor response range and optimization strategy.

[0030] Preferably, the parameter calibration module adopts the Brown-Conrady distortion model, which is expressed as: , where is the radius of the pixel after distortion, is the radius of the pixel before distortion, 、 、 is the radial distortion coefficient, which is calculated by photographing the calibration plate grid pattern and comparing the ideal grid with the distorted grid. 、 、 The image is remapped by fitting the distortion curve and the radius of the distorted pixel points;

[0031] The RGB gain calculation formula is: , where 、 、 is the gain value of RGB three channels, 、 、 When shooting a white reference plate, calculate 、 And applied to the R and B channels of the image to compensate for white balance deviations under different lighting conditions;

[0032] Color deviation Calculated using the CIEDE2000 color difference formula:

[0033] , where 、 、 is the Lab color value measured by the camera, 、 、 is the Lab reference value corresponding to the standard color card, and color correction is compensated using a 3×3 color mapping matrix: , where 、 、 is the color value of the original pixel, 、 、 is the corrected color value, are the color mapping matrix coefficients.

[0034] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0035] The present invention sets physical parameter thresholds based on the specifications of the camera module through a parameter calibration module, and then uses a calibration plate to calibrate the geometric and optical parameters to generate the boundary conditions of the camera module. The adaptive debugging module detects the environmental parameters of the shooting environment in real time and dynamically adjusts the imaging parameters of each camera. The multi-camera coordination module coordinates exposure across cameras and uses global white balance mapping to align curves according to the color temperature response curves of multiple cameras. Combined with a machine learning algorithm, it matches the color style of the color data of multiple cameras and adjusts the camera parameters when multiple cameras are switched. The debugging system adopts a layered debugging architecture combined with environmental adaptive optimization and cross-camera parameter coordination to achieve an automated, multi-camera collaborative high-precision parameter calibration process, thereby improving the consistency and stability of the camera module. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0037] Figure 1 It is a system module diagram of the present invention.

[0038] Figure 2 This is a mind map of the present invention.

[0039] Figure 3 This is the static structure class diagram of the debugging system of the present invention. DETAILED DESCRIPTION

[0040] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0041] Example 1: Please refer to Figure 1-Figure 3 As shown, the camera module automatic debugging system for adaptive parameter calibration described in this embodiment includes a parameter calibration module, an adaptive debugging module, and a multi-camera coordination module;

[0042] Parameter calibration module: During the factory debugging phase, physical parameter thresholds are set based on the camera module specifications. After geometric and optical parameter calibration using a calibration plate, the boundary conditions of the camera module are generated and sent to the adaptive debugging module and the adaptive debugging module.

[0043] Adaptive debugging module: This module detects the environmental parameters of the shooting environment in real time, including lighting, color temperature, and dynamic characteristics, and dynamically adjusts the imaging parameters of each camera to ensure stable imaging quality. The dynamically adjusted imaging parameters are sent to the multi-camera coordination module.

[0044] Multi-camera coordination module: This module coordinates exposure across cameras and uses global white balance mapping. It aligns the color temperature response curves of multiple cameras to avoid color jumps when switching between them. Combined with machine learning algorithms, it matches the color style of color data from multiple cameras to ensure consistent tones output by all cameras. It also adjusts camera parameters, including exposure, color, and distortion, when switching between multiple cameras (such as from the main camera to the telephoto lens or from the main camera to the ultra-wide angle lens), to ensure seamless transitions.

[0045] This application uses a parameter calibration module to set physical parameter thresholds based on the specifications of the camera module, and then uses a calibration plate to calibrate the geometric and optical parameters to generate the boundary conditions of the camera module. The adaptive debugging module detects the environmental parameters of the shooting environment in real time and dynamically adjusts the imaging parameters of each camera. The multi-camera coordination module coordinates exposure across cameras and uses global white balance mapping to align curves based on the color temperature response curves of multiple cameras. Combined with machine learning algorithms, it matches the color style of the color data of multiple cameras and adjusts the camera parameters when switching between multiple cameras. The debugging system uses a layered debugging architecture combined with environmental adaptive optimization and cross-camera parameter coordination to achieve an automated, multi-camera collaborative high-precision parameter calibration process to improve the consistency and stability of the camera module.

[0046] The specific workflow of the debugging system is as follows:

[0047] During the factory debugging stage, the definition layer sets the physical parameter thresholds based on the specifications of the camera module, and then uses a calibration plate to calibrate the geometric and optical parameters to generate the boundary conditions of the camera module. The debugging layer detects the environmental parameters of the shooting environment in real time. The environmental parameters include lighting, color temperature and dynamic characteristics, and dynamically adjusts the imaging parameters of each camera to ensure stable imaging quality. The coordination layer coordinates exposure across cameras and uses global white balance mapping. It aligns curves according to the color temperature response curves of multiple cameras to avoid color jumps when switching between multiple cameras. Combined with machine learning algorithms, the color data of multiple cameras is matched with color styles to ensure that the output tones of all cameras are consistent. When switching between multiple cameras (such as main camera to telephoto, main camera to ultra-wide angle), the camera parameters are adjusted. The camera parameters include exposure, color, and distortion to ensure seamless connection.

[0048] This application adopts a three-layer debugging architecture (core parameter calibration, dynamic coordination, and environmental adaptation), and each layer cooperates with each other to achieve an automated, multi-camera collaborative high-precision parameter calibration process.

[0049] The core parameter calibration layer mainly performs basic calibration on the optical characteristics and sensor parameters of the camera module to ensure that the physical performance of each camera meets the standards and provides boundary conditions for subsequent dynamic optimization.

[0050] During the factory commissioning phase, the system sets physical parameter thresholds based on the camera module specifications, including but not limited to:

[0051] Main camera (wide-angle): maximum aperture F1.8, supports PDAF (phase detection autofocus) and OIS (optical image stabilization).

[0052] Ultra wide angle: maximum distortion compensation Δd ≤ 2.5%, minimum focusing distance 10cm.

[0053] Telephoto: Minimum focusing distance 1.2m, equivalent focal length over 50mm, supports EIS (electronic image stabilization).

[0054] Calibration targets are used to calibrate geometric and optical parameters to ensure the basic imaging quality of the optical system. Lens focal length deviation is tested to ensure accurate optical magnification. Barrel distortion and pincushion distortion are corrected for the ultra-wide-angle camera. Color calibration is performed using a standard color chart to reduce color shift errors between different camera sensors. After calibration, the camera module enters the environmental adaptation debugging phase.

[0055] Environmental adaptive debugging layer: Real-time detection of the lighting, color temperature and dynamic characteristics of the shooting environment, and dynamic adjustment of the imaging parameters of each camera to ensure stable imaging quality.

[0056] Using sensor data and image analysis technology, the camera collects and analyzes external environmental information in real time. This includes determining whether the current shooting scene is in bright, low, or backlit conditions through light sensor and image histogram analysis. It also calculates color temperature distribution (Kelvin value), identifies the light source type (such as daylight, fluorescent light, or tungsten light), and applies the appropriate white balance calibration strategy. It also uses an optical flow algorithm to calculate the speed of moving objects within the scene and, combined with gyroscope data, determines whether optical or electronic image stabilization (OIS) is required.

[0057] Single-camera adaptive adjustment: Dynamically adjusts key parameters of each camera based on environmental data. For example, in low-light conditions, it increases exposure time or ISO to optimize brightness. In high-dynamic-range scenes (such as backlit scenes), it automatically activates HDR shooting mode to ensure balanced exposure between bright and dark areas. When device shake is detected, the telephoto lens automatically activates OIS (optical image stabilization) or EIS (electronic image stabilization).

[0058] Multi-camera parameter coordination balances hardware parameter limitations and environmental feedback to resolve parameter conflicts across multiple cameras.

[0059] Cross-camera exposure coordination: When the main camera narrows its aperture due to strong light, the B layer automatically increases the ISO of the telephoto camera to compensate for the amount of light entering, ensuring consistent brightness across multiple cameras. In Night mode, the B layer simultaneously adjusts the exposure time of the main and telephoto cameras to match the brightness curves of all cameras.

[0060] Global white balance mapping is used to align the color temperature response curves of multiple cameras to avoid color jumps when switching between them. Incorporating machine learning algorithms, color style matching is performed on the color data from multiple cameras to ensure consistent tones across all cameras. When switching between multiple cameras (e.g., from main camera to telephoto, or from main camera to ultra-wide angle), exposure, color, distortion, and other parameters are adjusted to ensure a seamless transition.

[0061] Traditional multi-camera systems rely on fixed parameter tables. This application implements intelligent, coordinated multi-camera tuning through dynamic parameter coordination (Layer B). By combining ambient lighting, color temperature, and dynamic information, camera parameters are adjusted in real time, enhancing imaging stability in complex lighting environments. Global white balance curve mapping eliminates multi-camera color shift and improves imaging consistency.

[0062] This application achieves high-precision automated debugging of camera modules through a three-layer architecture of core parameter calibration, environmental adaptive optimization, and multi-camera coordination, ensuring that the multi-camera system can maintain consistency and high-quality imaging effects in various complex scenarios, thereby improving user experience and device imaging capabilities.

[0063] Example 2: During the factory debugging phase, the parameter calibration module sets the physical parameter thresholds based on the specifications of the camera module, and then uses a calibration board to calibrate the geometric and optical parameters to generate the boundary conditions of the camera module. The boundary conditions are sent to the adaptive debugging module and the adaptive debugging module.

[0064] The parameter calibration module sets thresholds for key physical parameters based on the optical specifications and hardware characteristics of the camera module, including:

[0065] Aperture size: For example, if the maximum aperture of the main camera is set to F1.8, the minimum focusing distance of the telephoto lens is set to 1.2m.

[0066] Focal length range: Set the effective focal length range of wide-angle, ultra-wide-angle, and telephoto lenses based on the optical design of different cameras.

[0067] Distortion Compensation Limit: For ultra-wide-angle cameras, set the allowable error range for Barre-Distortion and Pincushion-Distortion (e.g., Δd ≤ 2.5%).

[0068] Light Sensitivity: Set the ISO sensitivity range to ensure consistent performance across cameras in low-light environments.

[0069] To ensure that the camera's optical system meets the design requirements, a calibration plate is required to calibrate the geometric parameters, including:

[0070] (1) Focal length calibration

[0071] Using a high-precision calibration plate, we capture a standard calibration pattern at a fixed distance to verify that the camera's actual focal length matches the nominal focal length. We calculate the optical magnification and correct for focal length errors, ensuring consistent depth of field across different cameras in the same scene.

[0072] (2) Distortion correction

[0073] For ultra-wide-angle cameras, we capture a calibration grid pattern and analyze the degree of Barrel Distortion and Pincushion Distortion. We then use an optical distortion correction algorithm to calculate compensation parameters and optimize the lens correction curve to minimize distortion around the edges of the image.

[0074] Under different lighting conditions, the color response of camera sensors varies. To ensure color consistency across multiple cameras, perform the following optical calibration:

[0075] Under a standard color temperature light source (such as 3200K, 5500K, or 6500K), photograph a white reference plate and measure the response values ​​of the RGB channels. Calculate the white balance gain parameter (RGB ratio) to correct white balance deviations between different cameras, ensuring consistent color temperature response curves.

[0076] Use a standard color chart (such as the X-Rite Color Checker) to perform color comparisons and analyze the camera's color reproduction capabilities under different light sources. Calculate color differences (ΔE values) and adjust the color mapping matrix to ensure consistent color performance across different cameras.

[0077] The parameter calibration module adopts the Brown-Conrady distortion model. The radial distortion expression of the Brown-Conrady distortion model is as follows: , where is the radius of the pixel after distortion (the image coordinate after distortion), is the radius of the pixel before distortion (ideal undistorted coordinate), 、 、 is the radial distortion coefficient, which is obtained through calibration calculation. By shooting the calibration plate grid pattern and comparing the ideal grid with the distorted grid, the 、 、 The image is remapped using the radius of the distorted pixels by fitting the distortion curve (this step is prior art and will not be described in detail in this application). The Brown-Conrady distortion model is used to describe the degree of distortion in wide-angle lenses, and the distortion correction parameters are calculated by solving it. After distortion correction, the image grid deformation is corrected, and straight lines remain straight.

[0078] The purpose of white balance correction is to adjust the gain of the RGB channels so that white objects still appear neutral white in different color temperature environments. The RGB gain calculation formula is as follows:

[0079] , where 、 、 is the gain value of RGB three channels, 、 、 The measurement value of the RGB sensor when shooting a white reference plate, with the G channel as the reference channel (set ), by calculating 、 Applied to the R and B channels of the image, it can compensate for white balance deviations under different lighting conditions so that white objects will not appear color cast. For example, in a 3200K (warm light) environment, the R channel is too high and the B channel is too low. Color balance can be achieved by adjusting the gain.

[0080] In order to make the camera's color reproduction more accurate under different light conditions, color mapping correction is required to correct color deviation. The CIEDE2000 color difference formula is usually used for calculation: , where 、 、 is the Lab color value measured by the camera, 、 、 The Lab reference value corresponding to the standard color card, color correction can be compensated using a 3×3 color mapping matrix: , where 、 、 is the color value of the original pixel, 、 、 is the corrected color value, For color mapping matrix coefficients, obtain the RGB color value of the camera by shooting a standard color card (such as X-Rite Color Checker), calculate the Lab color value of the standard color card, and convert it into RGB reference value, and solve it by least squares fitting method. Matrix (this step belongs to the existing technology and will not be described in detail in this application) to make the RGB output of the camera as close to the standard value as possible and the color deviation Used to measure color deviation and ensure the accuracy of camera color reproduction, the color mapping matrix optimizes color consistency and eliminates color deviation by adjusting the ratio of the three RGB channels.

[0081] To improve the camera’s low-light performance and dynamic range, the following sensor optimizations are required:

[0082] Test the camera's ISO sensitivity range in low-light environments (light intensity below a threshold) to ensure acceptable noise levels as the ISO is increased. Set the lowest usable ISO for each camera to avoid image quality degradation and loss of detail.

[0083] Use a high-contrast test scene (contrast greater than the contrast threshold) to measure the camera's dynamic range (the range of bright and dark areas that can be recorded). Set the HDR trigger threshold to ensure that the camera automatically activates HDR in high-dynamic-range scenes, improving shadow detail.

[0084] After completing all the above calibrations, the system generates the boundary conditions of the camera module based on the test results, including: optical limitation parameters (such as maximum aperture, minimum focal length, distortion compensation range), color correction matrix (white balance gain, color mapping matrix), sensor response range (ISO sensitivity, dynamic range threshold), and optimization strategy (HDR trigger point, automatic exposure compensation range).

[0085] These boundary conditions will serve as the basis for environmental adaptive debugging and multi-camera parameter coordination, ensuring that the camera can intelligently adjust imaging parameters in various scenarios and achieve consistent optimization of multiple cameras.

[0086] The adaptive debugging module detects the environmental parameters of the shooting environment in real time, including lighting, color temperature and dynamic characteristics, and dynamically adjusts the imaging parameters of each camera to ensure stable imaging quality. The dynamically adjusted imaging parameters are sent to the multi-camera coordination module.

[0087] The adaptive debugging module detects light intensity, color temperature, and dynamic characteristics in real time to adapt to different scenarios (such as backlight, low light, and motion). It detects scene brightness through the exposure value of the RGB sensor or CMOS sensor and calculates the exposure brightness. The expression is: , where is the exposure brightness, is the average brightness value of the image (calculated as: Y=0.299R+0.587G+0.114B), If is the lens aperture area, then: if the exposure brightness is less than the minimum exposure brightness threshold, it is a low-light scene; if the exposure brightness is greater than the maximum exposure brightness threshold, it is a high-light scene; and if the exposure brightness is within the range between the minimum exposure brightness threshold and the maximum exposure brightness threshold, it is a normal-light scene.

[0088] The Gray-World-Assumption algorithm is used to calculate the scene color temperature. The expression is: , where is the estimated color temperature, 、 、 is the average value of the three channels of the image. If the estimated color temperature is greater than the maximum color temperature threshold, it indicates a warm light source. In this case, the R gain needs to be reduced and the B gain needs to be increased. If the estimated color temperature is less than the maximum color temperature threshold, it indicates a cool light source. In this case, the R gain needs to be increased and the B gain needs to be reduced. The speed of the moving object is calculated using the expression: , where is the object speed, is the displacement of the object in two consecutive frames, The frame interval is the time between frames. If the object speed is greater than the speed threshold, it is a high-speed motion scene. The exposure time is shortened and OIS / EIS anti-shake is enabled. If the object speed is less than the speed threshold, it is a static scene. The exposure time is appropriately extended to improve image quality.

[0089] Based on the environmental parameter detection results, the camera's exposure time, white balance, ISO, HDR, OIS and other parameters are adaptively adjusted. Exposure time adjustment: , where is the adjusted exposure time, is the exposure time before adjustment, is the exposure brightness, is the adjustment coefficient, Usually the value is 0.1, low light environment: increase To increase brightness, but need to cooperate with OIS to reduce shaking. High brightness environment: shorten To prevent overexposure and trigger HDR mode.

[0090] Calculate the new RGB gains: , where is the adjusted RGN gain, is the RGN gain before adjustment, is the standard color temperature, The RGB gain is adjusted in real time based on the currently detected color temperature to ensure a stable white balance.

[0091] Calculating ISO settings: , For the adjusted value, Before adjustment value, is the target brightness, For exposure brightness, in low-light scenes: increase ISO, but avoid excessive increase to increase noise; in bright light scenes: lower ISO to reduce overexposure. If a high-brightness contrast scene (such as backlight) is detected, enable HDR.

[0092] The multi-camera coordination module is used to coordinate exposure across cameras and adopts global white balance mapping. It aligns the color temperature response curves of multiple cameras to avoid color jumps when switching between multiple cameras. Combined with machine learning algorithms, it matches the color style of the color data of multiple cameras to ensure that the tones output by all cameras are consistent. It also adjusts the camera parameters when switching between multiple cameras (such as main camera to telephoto camera, main camera to ultra-wide angle camera), including exposure, color, and distortion, to ensure seamless connection.

[0093] When multiple cameras work together, different cameras may experience different exposures for the same scene (for example, the main camera lets in more light due to its large aperture, while the telephoto camera lets in less light). To ensure exposure consistency, an exposure conversion model is needed to keep the exposure values ​​of different cameras stable when switching.

[0094] The multi-camera coordination module coordinates exposure across cameras and converts exposure parameters. The expression is: , where is the exposure value after conversion of the target camera, is the exposure value of the reference camera (such as the main camera), is the aperture diameter of the reference camera, is the aperture diameter of the target camera, is the ISO sensitivity of the target camera, is the ISO sensitivity of the reference camera, is the exposure time of the target camera, Using the camera's exposure time as a reference, if the main camera's exposure is overexposed, the telephoto lens's ISO will be increased to match the incoming light. If the telephoto lens's sensitivity is low, the exposure time will be extended, and OIS will be used to compensate for the effects of image stabilization. If the ultra-wide-angle lens has a smaller aperture, the exposure time will need to be increased appropriately. Since ultra-wide-angle lenses are susceptible to distortion, HDR will need to be used to enhance dark details.

[0095] The photosensitive elements of different cameras respond differently to color temperature, which can easily lead to white balance offset. Therefore, a global white balance mapping must be established to ensure consistent white balance among multiple cameras in different lighting environments. The white balance gain calculation expression of the multi-camera coordination module is: , where is the white balance gain of the target camera, is the white balance gain of the reference camera (main camera), is the color temperature detected by the target camera, Using the color temperature detected by the reference camera as a reference, when the color temperature is high (warm), the R channel gain is reduced and the B channel gain is increased. When the color temperature is low (cool), the R channel gain is increased and the B channel gain is reduced. When multiple cameras are unified, the global color temperature average is calculated and white balance is synchronized for all cameras.

[0096] A machine learning model is used to analyze the color shift of different cameras under the same lighting conditions, and style unification is achieved based on color distribution histogram matching. The color histogram matching algorithm is: , where is the color adjustment value of the target camera, is the color value of the reference camera, 、 The color histogram features of the target camera and the reference camera are used to learn the color style based on samples. The color curves of different cameras are adjusted to ensure that their output color style is consistent with that of the main camera. For HDR and low-light scenes, color mapping is optimized separately to prevent color drift in dark areas.

[0097] Adjusting camera parameters when switching between multiple cameras (e.g., from main camera to telephoto, or from main camera to ultra-wide-angle) is crucial. Ultra-wide-angle cameras are prone to barrel distortion due to their wide field of view. Optimizing the distortion correction model gradually adjusts exposure parameters during camera switching to avoid sudden brightness jumps. Smoothly adjusts RGB gain to avoid sudden color jumps. In low-light environments, the color temperature adjustment is reduced to minimize color drift. This adjustment process is prior art and will not be further detailed in this application.

[0098] A machine learning model is used to analyze the color shift of different cameras under the same lighting conditions. To ensure color consistency among multiple cameras under the same lighting conditions, the multi-camera coordination module introduces a machine learning model to analyze the color shift of different cameras and automatically adjust it. The specific method includes the following steps:

[0099] Under laboratory lighting conditions, use a standard color chart (such as an X-Rite ColorChecker) to capture raw color data from different cameras under the same lighting. Record the color temperature (Kelvin), illuminance (Lux), and ambient spectral distribution during capture and store for later analysis. Have all cameras capture the same scene separately and record the RAW image data for subsequent color shift calculations.

[0100] Convert the RAW data of all cameras to a unified color space (such as CIE XYZ or Lab color space) for standardized comparison. Calculate the color mapping relationship of each camera under different lighting conditions: , where is the color offset matrix used to adjust the color of the target camera to the color style of the reference camera. is the color value of the reference camera on the standard color card, is the color value captured by the target camera in the same environment, It is a transformation function used to fit the color mapping relationship, which can be trained using a second-order polynomial transformation or a neural network.

[0101] Using supervised learning methods, a machine learning model is trained using labeled color data. Features are extracted, including color histograms, RGB value distributions, and white balance gain parameters. A neural network color mapping model (suitable for large datasets) is trained to predict color mapping relationships between different cameras.

[0102] The neural network (suitable for large-scale datasets) color mapping model structure is as follows:

[0103] The input dimensions of the input layer include: color histogram features, RGB value statistical features (mean, variance, etc.), white balance gain parameters, and environmental parameters (color temperature, illumination);

[0104] Hidden layer (deep learning network): Fully Connected Neural Network (FCNN) architecture: First layer: 128 neurons, ReLU activation; second layer: 64 neurons, ReLU activation; third layer: 32 neurons, ReLU activation; fourth layer: outputs color shift correction parameters;

[0105] Output layer, used to predict the color map matrix.

[0106] Use the test dataset to verify the model's accuracy and calculate the average color shift ΔE (CIEDE2000 standard). If ΔE is too large, adjust the hyperparameters and retrain the model until the color consistency requirements are met.

[0107] Compared to traditional LUT (Look-Up-Table) correction methods, this method dynamically adapts to varying lighting environments and offers greater generalization capabilities. It uses a data-driven approach to analyze inherent camera color shifts, achieving more consistent color styles across different cameras. Combined with environmental parameter detection, it enables intelligent color adjustment, ensuring color consistency across multiple cameras. This method utilizes machine learning to analyze color shifts across multiple cameras, constructing a color mapping model. This method then makes real-time adjustments to ensure color consistency across multiple cameras, avoiding color jumps when switching between cameras and improving image quality.

[0108] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0109] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0110] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application. Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0111] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. The camera module automatic debugging system with adaptive parameter calibration is characterized by: Including parameter calibration module, adaptive debugging module, and multi-camera coordination module; Parameter calibration module: During the factory commissioning phase, physical parameter thresholds are set based on the camera module's specifications. A calibration plate is then used to calibrate the geometric and optical parameters to generate the camera module's boundary conditions. Adaptive debugging module: real-time detection of environmental parameters of the shooting environment and dynamic adjustment of imaging parameters of each camera; Multi-camera coordination module: This module coordinates exposure across cameras, uses global white balance mapping, aligns the color temperature response curves of multiple cameras, and combines machine learning algorithms to perform color style matching on color data from multiple cameras. The multi-camera coordination module coordinates exposure across cameras and converts exposure parameters. The expression is: , where is the exposure value after conversion of the target camera, is the exposure value of the reference camera, is the aperture diameter of the reference camera, is the aperture diameter of the target camera, is the ISO sensitivity of the target camera, is the ISO sensitivity of the reference camera, is the exposure time of the target camera, is the exposure time of the reference camera.

2. The camera module automatic debugging system with adaptive parameter calibration according to claim 1, characterized in that: The multi-camera coordination module calculates the white balance gain, and the expression is: , where is the white balance gain of the target camera, is the white balance gain of the reference camera, is the color temperature detected by the target camera, The color temperature detected by the reference camera.

3. The camera module automatic debugging system with adaptive parameter calibration according to claim 2, characterized in that: The multi-camera coordination module uses a machine learning model to analyze the color shift of different cameras under the same lighting conditions and unifies the style based on color distribution histogram matching. The color histogram matching algorithm is: , where is the color adjustment value of the target camera, is the color value of the reference camera, 、 is the color histogram feature of the target camera and the reference camera.

4. The camera module automatic debugging system with adaptive parameter calibration according to claim 3, characterized in that: The adaptive debugging module adaptively adjusts the camera's imaging parameters and exposure time based on the environmental parameter detection results: , where is the adjusted exposure time, is the exposure time before adjustment, is the exposure brightness, is the adjustment coefficient; Calculate the new RGB gains: , where is the adjusted RGN gain, is the RGN gain before adjustment, is the standard color temperature, is the currently detected color temperature; Calculating ISO settings: , For the adjusted value, Before adjustment value, is the target brightness, is the exposure brightness.

5. The camera module automatic debugging system with adaptive parameter calibration according to claim 4, characterized in that: The adaptive debugging module detects the scene brightness through the exposure value of the RGB sensor or the CMOS sensor and calculates the exposure brightness. The expression is: , where is the exposure brightness, is the average brightness value of the image, is the lens aperture area; The gray world algorithm is used to calculate the scene color temperature, and the expression is: , where is the estimated color temperature, 、 、 is the average value of the three channels of the image; Calculate the speed of a moving object using the expression: , where is the object speed, is the displacement of the object in two consecutive frames, is the frame interval time.

6. The camera module automatic debugging system with adaptive parameter calibration according to claim 5, characterized in that: The parameter calibration module sets the thresholds of physical parameters, including aperture size, focal length range, distortion compensation limit, and photosensitivity, according to the optical specifications and hardware characteristics of the camera module; Use a calibration plate to perform geometric parameter calibration, including focal length calibration and distortion correction.

7. The camera module automatic debugging system with adaptive parameter calibration according to claim 6, characterized in that: The parameter calibration module shoots the calibration plate grid pattern with the ultra-wide-angle camera, analyzes the degree of deformation of Barrel-Distortion and Pincushion-Distortion, uses the optical distortion correction algorithm, calculates the compensation parameters, and optimizes the lens correction curve; Under a standard color temperature light source, photograph a white reference plate, measure the response values ​​of the RGB channels, calculate the white balance gain parameters, and correct the white balance deviation of different cameras; Use standard color cards for color comparison, analyze the camera's color reproduction capabilities under different light sources, calculate color difference, and adjust the color mapping matrix.

8. The camera module automatic debugging system with adaptive parameter calibration according to claim 7, characterized in that: The parameter calibration module tests the ISO sensitivity range of the camera in a low-light environment and sets the lowest available ISO for different cameras; Use high-contrast test scenes to measure the dynamic range of the camera and set the HDR trigger threshold so that the camera automatically turns on HDR in high dynamic range scenes; Generate the boundary conditions of the camera module based on the test results, including optical limit parameters, color correction matrix, sensor response range and optimization strategy.

9. The camera module automatic debugging system with adaptive parameter calibration according to claim 8, characterized in that: The parameter calibration module adopts the Brown-Conrady distortion model, which is expressed as: , where is the radius of the pixel after distortion, is the radius of the pixel before distortion, 、 、 is the radial distortion coefficient, which is calculated by photographing the calibration plate grid pattern and comparing the ideal grid with the distorted grid. 、 、 The image is remapped by fitting the distortion curve and the radius of the distorted pixel points; The RGB gain calculation formula is: , where 、 、 is the gain value of RGB three channels, 、 、 When shooting a white reference plate, calculate 、 And applied to the R and B channels of the image to compensate for white balance deviations under different lighting conditions; Color deviation Calculated using the CIEDE2000 color difference formula: , where 、 、 is the Lab color value measured by the camera, 、 、 is the Lab reference value corresponding to the standard color card, and color correction is compensated using a 3×3 color mapping matrix: , where 、 、 is the color value of the original pixel, 、 、 is the corrected color value, are the color mapping matrix coefficients.

Images