A global control method for visible light imaging parameters of a distributed photoelectric system

Abstract

A kind of distributed optoelectronic system visible light imaging parameter global control method, this method belongs to distributed optoelectronic detection imaging technical field.To generate the panorama image of global brightness, contrast consistent smooth transition, different aperture visible light imaging parameter needs to be dynamically adjusted and controlled by global control algorithm, global control method first needs to design global control overall framework, to realize global perception, joint optimization, parameter synchronous distribution.First, the imaging parameter of global control needs to be calibrated by laboratory radiance consistency to eliminate the inherent noise and difference caused by detector production process.Second, local-global two-stage joint photometry calculates global brightness statistical value, and average brightness deviation can be generated from preset brightness value expectation.Finally, the next frame exposure parameter is calculated by PD 2 control algorithm E i Based on this, exposure time and gain are obtained, and are issued to imaging unit by high-precision synchronous data channel to generate globally uniform real-time image.

Description

Technical Field

[0001] This invention belongs to the field of distributed photoelectric detection and imaging technology, specifically relating to a global control method for visible light imaging parameters of a distributed photoelectric system. Background Technology

[0002] Existing typical distributed optoelectronic systems generally construct an all-around sensing space through a carefully arranged set of optoelectronic sensors (on vehicles, aircraft), employing various image / signal processing algorithms to achieve situational awareness under complex weather conditions. Existing system solutions include... Figure 1 As shown in the figure. In this system, due to the different imaging conditions (such as zenith angle, target background, etc.) of the visible light cameras installed at different locations, there are significant differences in the brightness, contrast, and color of the output images from different cameras.

[0003] The current method of independently controlling and outputting low-bit-width (8-bit) image data from each visible light camera greatly increases the difficulty of image stitching, resulting in serious stitching marks, sudden brightness changes, ghosting, and other problems.

[0004] Another approach involves employing localized automatic exposure and automatic gain control for each visible light camera in a distributed optoelectronic system. Each camera node adjusts its parameters based on local metering results and histogram statistics. While this method ensures good image output from each camera, it results in poor system-wide visual consistency.

[0005] During the image stitching stage, local control strategies can lead to severe abrupt changes in brightness, chroma, and color saturation in overlapping areas of adjacent fields of view. Using consistency correction methods at the stitching seam stage often comes at the cost of reduced image dynamic range and signal-to-noise ratio, and is difficult to meet the requirements of visual observation under motion conditions.

[0006] Therefore, it is necessary to design a scheme that can integrate global image information statistics and improve the consistency of image brightness and contrast of each visible light camera node through a global synchronous control method of imaging parameters, so as to achieve global synchronous perception, global calculation and global optimization. Summary of the Invention

[0007] The technical problem this invention aims to solve is to overcome the differences in imaging conditions at different locations in a distributed imaging system for a motion platform, and to address the technical challenge of excessive differences in brightness, color, contrast, and dynamic range of the output images from cameras at different locations, which makes it difficult to stitch images together and output real-time images that fail to meet visual observation requirements. This invention proposes a global control method for visible light imaging parameters in a distributed optoelectronic system, establishing a high-precision synchronous execution mechanism, designing a centralized imaging parameter adjustment framework, and using global-local two-level joint metering and a global automatic exposure control algorithm to eliminate exposure parameter differences caused by differences in imaging conditions, thus providing uniform and consistent real-time images for image stitching.

[0008] This invention designs a globally perceptive hardware platform to collect raw data from different visible light cameras and perform global brightness information statistical calculations. This generates exposure times and analog / digital gain parameters for cameras at different locations, achieving consistency correction of wide-bit raw data. Finally, an image enhancement algorithm generates highly consistent image data, improving the visual observation effect of the stitched image. The technical solution adopted by this invention is as follows:

[0009] A method for global control of visible light imaging parameters in a distributed optoelectronic system includes the following steps:

[0010] (1) Overall framework design of global control

[0011] The overall control framework is designed for global perception, computational optimization, and synchronous distribution. The overall control framework includes a system global control unit and a centralized global controller. The system global control unit uses nodes of multiple visible light cameras as execution ends, and the centralized global controller manages the exposure and gain parameters of all visible light nodes.

[0012] (2) Radiometric consistency correction

[0013] A radiometric consistency correction model is introduced. By calibrating each visible light camera in the laboratory using a standard light source and reflector, a relative gain compensation table is generated, and a globally recommended exposure value is provided. The actual exposure parameters sent by the i-th visible light camera node are obtained. :

[0014]

[0015] Among them, E g E is the globally recommended exposure value. i The actual exposure parameters sent by the i-th visible light camera node. Let f be the calibration coefficient for the i-th visible light camera node, and f be the compensation mapping function determined by the radiometric consistency correction model.

[0016] (3) Local-global two-level joint metering

[0017] By collecting local and global metering vector statistics from each visible light camera node, and using a stitched overlapping field-of-view netlist, different weights are assigned to grids in overlapping and non-overlapping regions, with overlapping regions receiving a greater weight than non-overlapping regions. Then, a weighted average of the brightness values ​​from all grids is calculated to determine the global weighted average brightness. ;

[0018] (4) Global automatic exposure control algorithm

[0019] The system's global control unit uses PID control theory to calculate the global target grayscale value and the current global weighted average brightness deviation e(k):

[0020]

[0021] Where e(k) is the brightness control deviation of the k-th frame, and G is the current global weighted average brightness. The preset brightness expectation value;

[0022] Through PD 2 Incrementally calculate the adjustment amount for the next frame:

[0023]

[0024] Where, ΔE v Adjust the exposure for the next frame. , , These are the proportional coefficient, integral coefficient, and differential coefficient, respectively.

[0025] (5) Exposure parameter calculation and synchronous distribution

[0026] according to Calculate and obtain the global recommended exposure value E for the next frame. g Based on step (2), the actual exposure parameters sent by the i-th visible light camera node are calculated. Then based on the exposure parameters The exposure time and gain of the camera are calculated and sent to each camera node through a high-precision control channel.

[0027] Furthermore, the method for establishing the stitched overlapping field of view netlist in step (3) is to calibrate the intrinsic and extrinsic parameters of each visible light camera by using a fixed target method, and finally design and construct the stitched overlapping field of view netlist.

[0028] Furthermore, the local photometric vector in step (3) is the basic information used to extract the current camera statistical information for each visible light camera node in real time, including the average brightness of multiple regions and histogram statistical information; the global photometric vector is the reference information.

[0029] This invention, within a distributed optoelectronic system for a motion platform, designs a global control framework for visible light imaging parameters. It employs a centrally controlled, high-precision time synchronization mechanism to trigger exposure and imaging from cameras at different positions according to a specific frame rate. Through the design of local-global two-level joint metering, global automatic exposure control, and radiometric consistency correction, it dynamically generates real-time parameters (exposure time, gain, etc.) adapted to visible light cameras at different positions. This solves the technical challenges of severe stitching artifacts, abrupt brightness changes, and overlapping ghosting on high-speed platforms caused by significant differences in the output images and dynamic ranges of visible light cameras at different positions. Compared with existing technologies, it achieves the following technical advantages:

[0030] (1) The imaging parameters of each camera are globally controlled, resulting in high consistency in image brightness, contrast, and dynamic range. The image stitching and seam transitions are smooth, with no ghosting or trailing.

[0031] (2) High real-time performance and high time registration accuracy. An edge-center hardware control platform is adopted, and exposure and control are completed through a high-precision external synchronization triggering method, with time registration accuracy at the microsecond level.

[0032] (3) It has good versatility and is suitable for multi-spectral distributed optoelectronic systems. The global control method for visible light imaging parameters can be applied to multi-spectral composite optoelectronic systems such as infrared, ultraviolet, and shortwave.

[0033] (4) Low latency and high reliability. The global control hardware platform is designed and implemented using FPGA. Through asynchronous pipelined processing across clock domains, it achieves reliable control of multiple nodes and low-latency image processing. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of a typical distributed optoelectronic system in the background technology of this invention.

[0035] Figure 2 This is a flowchart of the global control method for visible light imaging parameters of the distributed optoelectronic system of the present invention.

[0036] Figure 3 This is a schematic diagram of the edge-center hardware control platform of the present invention.

[0037] Figure 4 This is a schematic diagram of the local-global photometric region of the present invention. Detailed Implementation

[0038] The technical solution of the present invention will be further described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0039] In this embodiment, the system is as follows: Figure 1As shown, the visible light cameras are mounted at different positions on the motion platform; the illustration indicates that six cameras are mounted around the motion carrier. In reality, the camera mounting positions are not necessarily symmetrical about the carrier. The edge-center hardware control platform design scheme is as follows: Figure 3 As shown, the core device of the control unit is an FPGA, which is interconnected with each visible light detector through a high-speed data bus (such as optical fiber), a differential synchronization bus, and a low-speed data transmission bus, providing a data, synchronization, and control hardware platform.

[0040] according to Figure 1 The ideal installation position of the camera is shown. The intrinsic and extrinsic parameters of each visible light camera are calibrated by using a fixed target, and a stitched and overlapping field of view netlist is constructed. Figure 4 This diagram illustrates the metering areas. The left side represents the metering area of ​​visible light detector A, and the right side represents the metering area of ​​visible light detector B. The area between them is the overlapping area, based on which metering is performed. The RAW data output from the detectors is divided into N+M grids, and the average brightness I and local histograms for each grid are calculated. Each camera sends the grid average brightness data and local histograms to the central control unit. The central control unit then generates the exposure time. and gain .

[0041] The edge-to-center hardware control platform generates high-precision hardware synchronization signals, which are then sent to each camera detector via a low-speed data bus. Each detector generates raw data, which is then pipelined across clock cycles to the central FPGA unit. Figure 3 As shown, after receiving the original images from each camera, the FPGA performs image enhancement algorithms. The enhancement algorithm consists of two parts: a "multi-camera image data processing link" and an "exposure parameter closed-loop control link." The "multi-camera image data processing link" includes multiple visible light camera nodes simultaneously acquiring images. The image data sequentially undergoes data merging, metering calculation, color interpolation / correction, spatial filtering, gamma correction, edge enhancement, and image output. The simultaneous acquisition of images by multiple visible light camera nodes, followed by data merging and metering calculation, forms the "exposure parameter closed-loop control link." In this link, the metering algorithm calculates exposure control parameters based on the brightness statistics of the merged images and feeds these parameters back to each visible light camera node, thus forming a frame-by-frame closed-loop adjustment.

[0042] The following provides a detailed description of the implementation of this invention, explaining its specific implementation method, purpose, and effects.

[0043] See appendix Figure 2 The present invention provides a global control method for visible light imaging parameters of a distributed optoelectronic system, comprising the following steps:

[0044] (1) Overall framework design of global control

[0045] The purpose of this step is to establish a centralized control architecture to achieve global perception, joint optimization, and synchronous parameter distribution, overcoming the output inconsistency problem caused by independent control of each camera. A general control framework is designed for global perception, computational optimization, and synchronous distribution. This framework includes a system global control unit and a centralized global controller. The system global control unit uses nodes from multiple visible light cameras as execution ends, and the centralized global controller manages the exposure and gain parameters of all visible light nodes. This edge-center collaborative architecture enables microsecond-level time synchronization and low-latency control, providing a hardware foundation for subsequent global parameter optimization.

[0046] (2) Radiometric consistency correction

[0047] The purpose of this step is to eliminate the inherent output differences between different camera detectors caused by factors such as manufacturing processes and nonlinear device responses, so that each camera outputs a consistent grayscale value under the same irradiance, providing accurate pre-calibration for global brightness control.

[0048] In practical implementation, a radiometric consistency correction model is introduced. Each visible light camera is calibrated in the laboratory using a standard light source and reflector, generating a relative gain compensation table and globally recommending exposure values. The actual exposure parameters sent by the visible light camera node are obtained. :

[0049]

[0050] Among them, E g E is the globally recommended exposure value. i These are the exposure parameters actually sent by the i-th visible light camera node. is the calibration coefficient for the i-th visible light camera node, and f is the compensation mapping function determined by the radiometric consistency correction model; this step ensures the consistency of the response of different cameras to the same scene under various exposure conditions, providing an accurate basis for subsequent global automatic exposure control.

[0051] (3) Local-global two-level joint metering

[0052] The purpose of this step is to integrate local and global scene information from each camera and accurately calculate the overall average brightness of the system using a weighted approach for overlapping regions, thus avoiding global parameter imbalances caused by metering deviations from a single camera. By collecting local and global metering vector statistics from each visible light camera node, and using a stitched overlapping field-of-view netlist, different weights are assigned to grids in overlapping and non-overlapping regions, with overlapping regions receiving a greater weight than non-overlapping regions. Finally, a weighted average of the brightness values ​​from all grids is calculated to determine the global weighted average brightness. The local metering vector is the basic information, which is used to extract the current camera statistics for each visible light camera node in real time, including the average brightness of multiple regions and histogram statistics; the global metering vector is the reference information.

[0053] Specifically, the intrinsic parameters (focal length, principal point, distortion coefficients) and extrinsic parameters (rotation matrix, translation vector) of each camera are first obtained through a fixed target calibration method. Then, based on the installation geometry and field of view, the overlapping field of view between adjacent cameras is calculated, establishing a stitched overlapping field of view netlist. This netlist records the camera number and overlap weight of each pixel location. During metering, each camera node divides the current frame RAW image into N×M grids (e.g., 8×6 grids), and calculates the average brightness value and histogram distribution of each grid as a local metering vector. Each node uploads its local metering vector to the central controller via a high-speed data bus. The central controller, based on the overlapping field of view netlist, assigns higher weights (e.g., weight 2) to grids in the overlapping region and basic weights (e.g., weight 1) to non-overlapping regions. Then, it performs a weighted average of the brightness values ​​of all grids to obtain the global weighted average brightness G. Simultaneously, the central controller also calculates a global histogram as an auxiliary reference. This two-stage combined metering method takes into account both local details and global consistency, effectively preventing excessive impact of local overexposure or underexposure on global statistics.

[0054] (4) Global automatic exposure control algorithm

[0055] The purpose of this step is to dynamically adjust the recommended global exposure value for the next frame based on the deviation between the global brightness statistics and the preset target, so that the system output brightness quickly converges to the desired value and remains stable. The system's global control unit uses PID control theory to calculate the deviation e(k) between the global target grayscale value and the current global weighted average brightness:

[0056]

[0057] Where e(k) is the brightness control deviation of the k-th frame, and G is the current global weighted average brightness. The preset brightness expectation value;

[0058] Through PD2 Incrementally calculate the adjustment amount for the next frame:

[0059]

[0060] Where, ΔE v Adjust the exposure for the next frame. , , These are the proportional coefficient, integral coefficient, and derivative coefficient; the values ​​of these three coefficients are preset based on the system frame rate and detector response characteristics, and adjusted through simulation or actual measurement to obtain the optimal convergence curve. Compared with traditional PID, this algorithm enhances stability against sudden changes in illumination, making it particularly suitable for scenarios with rapidly changing motion platforms.

[0061] (5) Exposure parameter calculation and synchronous distribution

[0062] The purpose of this step is to convert the calculated global recommended exposure value into an executable exposure time and gain for each camera, and to ensure that all cameras apply the new parameters at the same time through a high-precision synchronization channel, thereby achieving seamless inter-frame switching. The central controller, based on... Calculate and obtain the global recommended exposure value E for the next frame. g The exposure parameters of the i-th visible light camera are calculated according to the formula in step (2). Then based on the exposure parameters The exposure time and gain of the camera are calculated and sent to each camera node through a high-precision control channel.

[0063] In this step, the central controller writes the decomposed parameters into the registers of each camera via a low-speed data bus, and simultaneously sends an external synchronization pulse via a differential synchronization bus. The rising edge of the pulse triggers all cameras to simultaneously load the new parameters and begin exposure at the start of the next frame. This synchronization mechanism has an accuracy of microseconds, ensuring that the images from multiple cameras are strictly aligned in time, avoiding inter-frame brightness jumps caused by inconsistent parameter update times.

[0064] The above embodiments are merely for understanding the technical solution of the present invention and are not intended to limit the scope of protection of the present invention. Any obvious adjustments and modifications made to the technical solution of the present invention that fall within the inventive concept should also fall within the scope of protection of the present invention.

Claims

1. A method for global control of visible light imaging parameters in a distributed optoelectronic system, characterized in that, Includes the following steps: (1) Overall framework design of global control The overall control framework is designed for global perception, computational optimization, and synchronous distribution. The overall control framework includes a system global control unit and a centralized global controller. (2) Radiometric consistency correction A radiometric consistency correction model is introduced. By calibrating each visible light camera in the laboratory using a standard light source and reflector, a relative gain compensation table is generated, and a globally recommended exposure value is provided. The actual exposure parameters sent by the i-th visible light camera node are obtained. : Among them, E g E is the globally recommended exposure value. i The actual exposure parameters sent by the i-th visible light camera node. Let f be the calibration coefficient for the i-th visible light camera node, and f be the compensation mapping function determined by the radiometric consistency correction model. (3) Local-global two-level joint metering By collecting local and global metering vector statistics from each visible light camera node, and using a stitched overlapping field-of-view netlist, different weights are assigned to grids in overlapping and non-overlapping regions, with overlapping regions receiving a greater weight than non-overlapping regions. Then, a weighted average of the brightness values ​​from all grids is calculated to determine the global weighted average brightness. ; (4) Global automatic exposure control algorithm The system's global control unit uses PID control theory to calculate the global target grayscale value and the current global weighted average brightness deviation e(k): Where e(k) is the brightness control deviation of the k-th frame, and G is the current global weighted average brightness. The preset brightness expectation value; Through PD 2 Incrementally calculate the adjustment amount for the next frame: Where, ΔE v Adjust the exposure amount for the next frame. , , These are the proportional coefficient, integral coefficient, and differential coefficient, respectively. (5) Exposure parameter calculation and synchronous distribution according to Calculate and obtain the global recommended exposure value E for the next frame. g Based on step (2), the actual exposure parameters sent by the i-th visible light camera node are calculated. Then based on the exposure parameters The exposure time and gain of the camera are calculated and sent to each camera node through a high-precision control channel.

2. The global control method for visible light imaging parameters of a distributed optoelectronic system according to claim 1, characterized in that, In step (1), the system global control unit uses the nodes of multiple visible light cameras as execution ends and manages the exposure and gain parameters of all visible light nodes through a centralized global controller.

3. The global control method for visible light imaging parameters of a distributed optoelectronic system according to claim 1, characterized in that, Step (3) The method for establishing the stitched overlapping field of view netlist is to use a fixed target method to calibrate the intrinsic and extrinsic parameters of each visible light camera, and finally design and construct the stitched overlapping field of view netlist.

4. The global control method for visible light imaging parameters of a distributed optoelectronic system according to claim 1, characterized in that, The local photometric vector in step (3) is the basic information used to extract current camera statistics for each visible light camera node in real time, including the average brightness of multiple regions and histogram statistics; the global photometric vector is the reference information.

Images