A driving method and device for a vehicle-mounted camera system

Abstract

The present invention belongs to the technical field of vehicle-mounted cameras, and specifically relates to a driving method and device for a vehicle-mounted camera system, comprising the following steps: S1, dynamically monitoring the working state and environmental conditions of the vehicle-mounted camera through an intelligent power management unit, adaptively adjusting the power supply strategy according to the monitoring results, and realizing power consumption optimization; S2, implementing a dynamic calibration process, using a gyroscope to measure the angular velocity and an accelerometer to measure the acceleration, and combining the vehicle dynamics model to calculate the vehicle posture parameters S3, applying an image stabilization algorithm, and calculating the image motion vector based on the previous frame image and the current frame image; S4, implementing encrypted transmission with cross-platform compatibility and secure data transmission protocols; S5, integrating fault diagnosis and self-repair functions, analyzing system data through a fault detection algorithm, and performing repair operations. It reduces unnecessary energy consumption, reduces the aging problem of electronic components caused by overheating, and prolongs the life of the camera.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle-mounted cameras, and in particular relates to a driving method and device for a vehicle-mounted camera system. Background Art

[0002] In the context of the rapid development of today's automobile industry, the vehicle camera system, as an important component of the realization of advanced driver assistance systems (ADAS) and autonomous driving technology, is facing unprecedented challenges and demands. Although many technologies have been applied to the field of vehicle cameras, there are still some problems and shortcomings that need to be solved: traditional vehicle camera systems often lack efficient energy management mechanisms, usually use fixed power supply, and cannot be dynamically adjusted according to the actual working status of the camera and environmental changes. This not only causes energy waste, but may also affect the life of camera components due to excessive heat.

[0003] After searching, the existing Chinese patent publication number: CN116095436A is an adaptive vehicle camera driving method, comprising the following steps: S1, traversing all I2C buses to detect whether there are devices on each bus; if there are devices, traversing the devices on the bus to confirm the type of the device; according to the type of the device, obtaining the parameters of the device; according to the obtained device parameters, configuring various parameters of the camera module; obtaining information about the camera module accessing the deserializer; according to the obtained link information, the camera driver automatically configures the connection parameters between the camera module and the intelligent assisted driving system; the camera driver drives the camera module to output image data to the intelligent assisted driving system according to the configuration and connection parameters of the camera module.

[0004] The cited patent documents also have the same problem, lacking an efficient energy management mechanism and usually using a fixed power supply that cannot be dynamically adjusted according to the actual working state of the camera and environmental changes. This not only causes energy waste, but may also affect the life of the camera components due to excessive heat. Summary of the invention

[0005] The present invention provides a driving method and device for a vehicle-mounted camera system to solve the problems raised in the background technology.

[0006] A driving method for a vehicle-mounted camera system is provided, comprising the following steps:

[0007] S1, dynamically monitors the working status and environmental conditions of the vehicle-mounted camera through the intelligent power management unit (PMU), and adaptively adjusts the power supply strategy according to the monitoring results to achieve power consumption optimization;

[0008] S2, implementing a dynamic calibration process, using the angular velocity measured by the gyroscope and the acceleration measured by the accelerometer, and calculating the vehicle posture parameters in combination with the vehicle dynamics model;

[0009] S3, applying an image stabilization algorithm to calculate an image motion vector based on the previous frame image and the current frame image;

[0010] S4, using cross-platform compatibility and secure data transmission protocols to implement encrypted transmission;

[0011] S5, integrates fault diagnosis and self-repair functions, analyzes system data through fault detection algorithms, and performs repair operations;

[0012] S6, through an integrated deep learning model, based on convolutional neural networks (CNN) and recurrent neural networks (RNN), achieves real-time scene understanding and behavior prediction of the vehicle's surrounding environment;

[0013] S7, uses multi-camera collaborative technology and combines stereo vision principles to implement 3D scene reconstruction;

[0014] S8, receiving a first operation signal of a first camera module identifier input by a user, wherein the first camera module is connected to the first deserializer through a first communication channel, and according to the first operation, determining a first target serializer driver and a first target camera driver that match the identifier through a matching algorithm in a driver module file corresponding to the first deserializer.

[0015] This solution is preferred. By integrating an intelligent power management unit (PMU), the working status and surrounding environmental conditions of the camera are monitored in real time, and the power supply strategy is adaptively adjusted to achieve optimal power consumption. In particular, a power consumption prediction model based on machine learning is introduced, and the long short-term memory network (LSTM) is used to analyze historical power consumption data and predict the power distribution strategy for future periods, so as to fine-tune energy use, reduce energy consumption and reduce the risk of overheating.

[0016] This solution is optimized to use gyroscope and accelerometer data combined with vehicle dynamics model for dynamic calibration to ensure that camera parameters are adjusted to the optimal state in real time. On this basis, the image stabilization algorithm is applied to process the image sequence through the extended Kalman filter (EKF), effectively reducing image blur and improving image stability.

[0017] This solution is preferably compatible with protocols of multiple communication interface standards (such as CAN, Ethernet, and Wi-Fi), realizes secure encrypted transmission of data, and adopts an adaptive selection algorithm to ensure efficient data exchange and enhance the information security of the system.

[0018] This solution is preferably combined with fault detection and self-repair functions, and the system operation data is analyzed by algorithm to automatically identify and repair faults to ensure stable and reliable operation of the system.

[0019] This solution is preferred, through the deep learning model combined with convolutional neural network (CNN) and recurrent neural network (RNN), real-time analysis of the vehicle's surrounding environment and behavior prediction. In addition, through multi-camera collaborative technology and stereo vision principles, 3D scene reconstruction is implemented to improve the accuracy of obstacle recognition and environmental perception, providing more accurate environmental information for autonomous driving.

[0020] A driving device for a vehicle-mounted camera system comprises an intelligent power management unit, a dynamic calibration module, an image stabilization processor, a data transmission controller and a fault processing unit.

[0021] Preferably, the driving device of the present invention further includes a deep learning processing unit and a multi-camera collaborative processing module connected together via electrical signal transmission.

[0022] Technical effects and advantages of the present invention:

[0023] The driving method of the vehicle camera system, the intelligent power management unit (PMU) optimizes power consumption according to actual working requirements and environmental conditions by dynamically adjusting the power supply strategy, effectively reducing unnecessary energy consumption, while reducing the aging of electronic components caused by overheating, and extending the overall service life of the camera system. The combination of the dynamic calibration module and the image stabilization processor ensures that the images captured by the camera are clear and without obvious distortion when the vehicle is moving and in complex environments, improving the reliability and accuracy of the visual system;

[0024] By integrating deep learning models, especially the combination of convolutional neural networks (CNN) and recurrent neural networks (RNN), the system can more accurately understand complex driving scenarios, predict the dynamics of the vehicle's surrounding behaviors, improve the ability to predict emergencies, and reduce the risk of traffic accidents;

[0025] The application of multi-camera collaborative processing and 3D visual reconstruction technology not only improves the accuracy of obstacle detection and distance estimation, but also provides richer environmental information for autonomous driving, supports more accurate path planning and obstacle avoidance strategies, and enhances the reliability and safety of the driving assistance system. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The present invention is a block diagram of a driving method of a vehicle-mounted camera system. DETAILED DESCRIPTION

[0027] The technical solution of the present invention will be described below in conjunction with the accompanying drawings in the embodiments of the present invention. Figure 1 A driving method of a vehicle-mounted camera system shown includes the following steps:

[0028] S1, dynamically monitors the working status and environmental conditions of the vehicle-mounted camera through the intelligent power management unit (PMU), and adaptively adjusts the power supply strategy according to the monitoring results to achieve power consumption optimization;

[0029] The function formula for power consumption optimization is: ,in is the current power consumption, Indicates the camera status. represents environmental factors;

[0030] S2, implements the dynamic calibration process and uses the gyroscope to measure the angular velocity and the acceleration measured by the accelerometer , combined with the vehicle dynamics model Calculate vehicle posture parameters and adjust camera angle accordingly and focal length , ensuring image stability;

[0031] S3, apply the image stabilization algorithm based on the previous frame image and the current frame image , calculate the image motion vector , using inverse motion compensation technology Compensate Eliminate image jitter;

[0032] S4, using cross-platform compatibility and secure data transmission protocols to implement encrypted transmission ,in Indicates the encrypted data. is the key, is the original data, is Advanced Encryption Standard;

[0033] S5, integrates fault diagnosis and self-repair functions, through fault detection algorithm Analyze system data , and perform repair operations ,in For repair strategies, is the fault type;

[0034] S6, through an integrated deep learning model, based on convolutional neural network (CNN) and recurrent neural network (RNN), achieves real-time scene understanding and behavior prediction of the vehicle's surrounding environment. ,in To predict the results, is an image sequence;

[0035] S7, using multi-camera collaborative technology, combined with stereo vision principles, implements 3D scene reconstruction and calculates depth maps stereo_match ,in and Left and right camera images respectively;

[0036] S8, receiving a first operation signal of a first camera module identifier input by a user , wherein the first camera module and the first deserializer communicate with each other through a first communication channel connection, according to the first operation, in the driver module file corresponding to the first deserializer Through matching algorithm Determine the first target serializer driver that matches the identifier and the first target camera driver , the driver module file Integrated with the first deserializer driver , multiple serializer driver sets and multiple camera driver collections .

[0037] In this embodiment, the intelligent power management unit further includes a power consumption prediction model based on machine learning, which is based on the historical power consumption data sequence. and current working mode , using long short-term memory network (LSTM) to predict future time windows Power distribution strategy within , to optimize energy use.

[0038] In this embodiment, the dynamic calibration process uses a particle swarm optimization algorithm (PSO) to quickly adjust the camera parameters, with the number of iterations per second not exceeding Second, the optimization objective function is the image clarity index to maximize the real-time and accuracy.

[0039] In this embodiment, the image stabilization algorithm is combined with an extended Kalman filter (EKF) to obtain the state equation and the observation equation , reducing the prediction error , improve the image stabilization effect, including is the state vector, is the state transfer matrix, is the control input, is the process noise, is the observation vector, is the observation matrix, is the observation noise.

[0040] In this embodiment, the cross-platform compatibility and secure data transmission protocol supports a variety of different communication interface standards, including but not limited to CAN, Ethernet, Wi-Fi, and the use of protocol adaptive selection algorithm is the compatibility scoring function.

[0041] In this embodiment, the deep learning model training process uses a reinforcement learning algorithm to optimize the objective function ,in To accumulate rewards, is the regularization factor, As model parameters, it can improve the accuracy of scene understanding and prediction.

[0042] In this embodiment, the multi-camera collaborative technology uses time synchronization and spatial registration algorithms to ensure accurate alignment between image sequences. , improving the accuracy of 3D reconstruction.

[0043] In this embodiment, in the vehicle operating system kernel Without pre-installed vehicle camera solidification driver, the driver module file Load into kernel , load the first deserializer driver , First target serializer driver And the first target camera driver , to activate the image transmission function between the first deserializer and the first camera module ;

[0044] Before determining the matching driver, perform the acquisition step to collect all deserializer drivers , Serializer Driver and camera driver .

[0045] A driving device for a vehicle-mounted camera system comprises an intelligent power management unit, a dynamic calibration module, an image stabilization processor, a data transmission controller and a fault processing unit. The driving device also comprises a deep learning processing unit and a multi-camera collaborative processing module connected together through electrical signal transmission.

[0046] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A driving method for a vehicle-mounted camera system, characterized in that: The following steps are involved: S1, dynamically monitors the working status and environmental conditions of the vehicle camera through the intelligent power management unit PMU, and adaptively adjusts the power supply strategy according to the monitoring results to achieve power consumption optimization; S2, implementing a dynamic calibration process, using the angular velocity measured by the gyroscope and the acceleration measured by the accelerometer, and calculating the vehicle posture parameters in combination with the vehicle dynamics model; S3, apply the image stabilization algorithm based on the previous frame image and the current frame image , calculate the image motion vector , Inverse Motion Compensation Technology Compensate Eliminate image jitter; S4, using cross-platform compatibility and secure data transmission protocols to implement encrypted transmission; S5, integrates fault diagnosis and self-repair functions, analyzes system data through fault detection algorithms, and performs repair operations; S6, through an integrated deep learning model, based on convolutional neural network (CNN) and recurrent neural network (RNN), achieves real-time scene understanding and behavior prediction of the vehicle's surrounding environment; S7, uses multi-camera collaborative technology and combines stereo vision principles to implement 3D scene reconstruction; S8, receiving a first operation signal of a first camera module identifier input by a user, wherein the first camera module is connected to the first deserializer through a first communication channel, and determining, according to the first operation signal, a first target serializer driver and a first target camera driver that match the identifier through a matching algorithm in a driver module file corresponding to the first deserializer; The intelligent power management unit further includes a power consumption prediction model based on machine learning, which is based on the historical power consumption data sequence. and current working mode , using the long short-term memory network LSTM to predict future time windows Power distribution strategy within , to optimize energy use; The dynamic calibration process uses the particle swarm optimization algorithm PSO to quickly adjust the camera parameters, with the number of iterations per second not exceeding Second, the optimization objective function is the image clarity index Maximization of data quality to ensure real-time and accuracy; The image stabilization algorithm combines the extended Kalman filter EKF through the state equation and the observation equation , reducing the prediction error , improve the image stabilization effect, including is the state vector, is the state transfer matrix, is the control input, is the process noise, is the observation vector, is the observation matrix, is the observation noise.

2. The driving method of a vehicle-mounted camera system according to claim 1, characterized in that: The cross-platform compatibility and secure data transmission protocol supports a variety of different communication interface standards, including but not limited to CAN, Ethernet, Wi-Fi, and adopts a protocol adaptive selection algorithm is the compatibility scoring function.

3. The driving method of a vehicle-mounted camera system according to claim 1, characterized in that: The deep learning model training process uses a reinforcement learning algorithm to optimize the objective function ,in To accumulate rewards, is the regularization factor, As model parameters, to improve the accuracy of scene understanding and prediction.

4. The driving method of a vehicle-mounted camera system according to claim 3, characterized in that: The multi-camera collaborative technology uses time synchronization and spatial registration algorithms to ensure accurate alignment between image sequences. , improving the accuracy of 3D reconstruction.

5. The driving method of a vehicle-mounted camera system according to claim 1, characterized in that: In the vehicle operating system kernel Without pre-installed vehicle camera fixed driver, the driver module file Load into kernel , load the first deserializer driver , First target serializer driver And the first target camera driver , to activate the image transmission function between the first deserializer and the first camera module ; Before determining the matching driver, perform the acquisition step to collect all deserializer drivers , Serializer Driver and camera driver .

6. A driving device obtained by using the driving method of a vehicle-mounted camera system according to any one of claims 1 to 5, characterized in that: It includes an intelligent power management unit, a dynamic calibration module, an image stabilization processor, a data transmission controller and a fault processing unit.

7. The driving device according to claim 6, characterized in that: The driving device also includes a deep learning processing unit and a multi-camera collaborative processing module connected together through electrical signal transmission.

Images