Automobile data recorder data security storage and traceability system fused with block chain technology

Abstract

The invention relates to an automobile data recorder data security storage and traceability system fused with a block chain technology, belongs to the field of internet platforms, and more particularly belongs to the field of accounting systems for registering or indicating operation of machines. The block chain encryption module is used for carrying out block chain encryption storage on each frame of driving picture recorded by the driving recorder; and the intelligent analysis equipment is used for intelligently analyzing whether a vehicle turning accident occurs in the current time interval or not by adopting an AI accident analysis model according to the plurality of first accounting values and the plurality of second accounting values respectively corresponding to the plurality of frames of passing driving pictures corresponding to the current moment. According to the invention, for the technical problem that the accounting data for accident early warning cannot be collected according to the road condition, the road condition visual data can be acquired by adopting a data storage and tracing mechanism based on the block chain, and then various related accounting data are analyzed and used for intelligently analyzing the occurrence of the vehicle turning accident, so that the technical problem is solved.

Description

Technical Field

[0001] The Internet platform of the present invention more specifically relates to an accounting system for registering or indicating the operation of a machine, and in particular to a dashcam data security storage and tracing system that integrates blockchain technology. Background Art

[0002] With the continuous upgrading of vehicle intelligent transformation, improving the controllability and safety of vehicle driving based on the Internet platform has become one of the main trends in vehicle design. The Internet platform can facilitate the intelligent transformation of vehicles in many aspects such as data security storage and traceability and data interaction. At the same time, vehicle designers hope to be able to use the accounting data collected according to the ever-changing road conditions during the vehicle driving process to determine whether the vehicle as a machine will be in an accident-prone state in the future, and then instruct the vehicle's operation to avoid future accidents as much as possible, thereby improving the controllability and safety of the vehicle.

[0003] For example, Chinese invention patent publication CN114970902A proposes a blockchain-based vehicle power battery carbon emissions accounting system. The system includes: a production process carbon emissions module, connected to a blockchain module, for recording carbon emissions during the production process and uploading the carbon emissions to the blockchain module; an operation process carbon emissions module, connected to a blockchain module, for recording carbon emissions during vehicle operation. This blockchain-based vehicle power battery carbon emissions accounting system can monitor the carbon emissions generated by the power battery during vehicle operation, as well as the carbon emissions generated during vehicle production and recycling, thereby obtaining the carbon emissions of the vehicle throughout its life cycle. The carbon emissions information is then shared with enterprises and regulatory authorities to facilitate subsequent optimization of the production process and reduce carbon emissions.

[0004] For example, Chinese invention patent publication CN108256766A proposes a vehicle insurance calculation method based on dangerous driving behavior. The method comprises: assessing the level of dangerous driving behavior; obtaining dangerous driving behavior information; obtaining vehicle information; and calculating vehicle insurance premiums by analyzing and grading the dangerous driving behavior information and combining it with the vehicle information. This invention collects and analyzes information on dangerous driving behavior and assigns a risk level weight to each dangerous driving behavior, thereby deriving a vehicle insurance calculation method based on dangerous driving behavior. This improves the rationality of the insurance calculation method and encourages drivers to reduce dangerous driving behavior, thereby reducing traffic accidents.

[0005] Obviously, the above-mentioned existing technologies are unable to collect various calculation data related to traffic accidents from the ever-changing road conditions during the vehicle's driving process, resulting in an inability to determine whether the vehicle as a machine will be in an accident-prone state in the future based on the various calculation data related to traffic accidents, and thus unable to instruct the operation of the vehicle to avoid the occurrence of future accidents as much as possible, affecting the improvement of the vehicle's controllability and safety. At the same time, the driving data related to the ever-changing road conditions lacks a secure data storage and traceability mechanism, and cannot provide sufficiently reliable reference data for subsequent vehicle operation instructions. Summary of the Invention

[0006] To address the technical problems in the prior art, the present invention provides a dashcam data security storage and tracing system that integrates blockchain technology. This system is capable of performing blockchain-encrypted storage on each frame of driving footage recorded by the dashcam during vehicle driving. Each frame of driving footage is assigned a blockchain storage address in a distributed ledger storage mode for performing blockchain-encrypted storage on that frame of driving footage. Blockchain-encrypted storage is also performed using a consensus mechanism, thereby providing a data security storage mechanism for subsequent tracing of multiple frames of driving footage corresponding to a set time. More importantly, the system also intelligently analyzes whether a vehicle turning accident will occur within a future time interval starting at the set time based on two different accounting values ​​obtained by analyzing various visual data of multiple frames of driving footage before the set time. If the intelligent analysis determines that a vehicle turning accident will occur within the future time interval starting at the set time, the system instructs the vehicle to enter a cautious operation mode. Otherwise, the system instructs the vehicle to exit the cautious operation mode, thereby establishing an accounting system that indicates the operation of the machine and improves the predictability and safety of driving.

[0007] According to the present invention, a driving recorder data security storage and tracing system integrating blockchain technology is provided, the system comprising: An encrypted storage device is used to encrypt and store each frame of driving footage recorded by the driving recorder during the vehicle's driving process in a blockchain. The driving recorder uses a set frame rate; A data tracing device is connected to the encrypted storage device and is used to trace back multiple frames of driving images before the current moment as multiple frames of past driving images corresponding to the current moment; a first calculation device, connected to the data tracing device, configured to uniformly divide the road surface imaging sub-image in each frame of the past driving image into equal areas to obtain uniformly divided blocks of the road surface imaging sub-image in the frame, and calculate the number of uniformly divided blocks in the road surface imaging sub-image in the frame whose component standard deviation exceeds a limit as a first calculation value corresponding to the frame of the past driving image; A second calculation device, connected to the first calculation device, is used to calculate the number of curves in the road imaging sub-frame of each frame of the driving picture whose curvature is greater than or equal to a preset curvature value as a second calculation value corresponding to each frame of the driving picture; The intelligent analysis device is connected to the first calculation device and the second calculation device respectively, and is used to use the AI ​​accident analysis model to intelligently analyze the accident occurrence signs of vehicle turning accidents in the current time interval according to the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving pictures corresponding to the current moment, the multiple first calculation values ​​corresponding to the multiple frames of past driving pictures corresponding to the current moment, and the multiple second calculation values.

[0008] According to a second aspect of the present invention, a dashcam data security storage and tracing system integrating blockchain technology is provided. The system includes a memory and multiple processors. The memory stores a computer program. The computer program is configured to be executed by the multiple processors to complete the following steps: Each frame of driving footage recorded by the driving recorder during the vehicle's driving process is encrypted and stored on the blockchain. The driving recorder uses a set frame rate; At the current moment, multiple frames of driving images before the current moment are traced back to serve as multiple frames of past driving images corresponding to the current moment; Performing uniform segmentation of equal areas on the road imaging sub-image in each frame of the past driving image to obtain uniformly segmented blocks of the road imaging sub-image in the frame, and calculating the number of uniformly segmented blocks in the road imaging sub-image in the frame whose component standard deviation exceeds a limit as a first calculation value corresponding to the frame of the past driving image; Calculating the number of curves in the road imaging sub-frame of each frame of the driving picture whose curvature is greater than or equal to a preset curvature value as a second calculation value corresponding to each frame of the driving picture; An AI accident analysis model is used to intelligently analyze the accident occurrence signs of vehicle turning accidents in the current time interval based on the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first calculation values ​​corresponding to the multiple frames of the past driving images corresponding to the current moment, and the multiple second calculation values.

[0009] It can be seen that the present invention has at least the following outstanding substantive features: Substantive Feature 1: Each frame of driving footage recorded by the dashcam during driving is encrypted and stored on the blockchain. The dashcam uses a set frame rate, and each frame of driving footage is assigned a blockchain storage address in the blockchain distributed ledger storage mode for blockchain encrypted storage of that frame of driving footage. This is performed using a consensus mechanism, thereby providing a data security storage mechanism for the subsequent tracing of multiple frames of driving footage corresponding to a set time. Substantive Feature 2: Based on two different calculation values ​​obtained by analyzing various visual data of multiple frames of driving images before a set time, an intelligent analysis is performed to determine whether a vehicle turning accident will occur in a future time interval starting at the set time. If the intelligent analysis determines that a vehicle turning accident will occur in the future time interval starting at the set time, the vehicle is instructed to enter cautious operation mode. Otherwise, the vehicle is instructed to exit cautious operation mode, thereby establishing a calculation system that indicates the operation of the machine, improving the predictability and safety of driving; Substantive Feature Three: To achieve intelligent analysis of whether a vehicle turning accident will occur in a future time interval starting at a set time, a customized AI accident analysis model is introduced. The AI ​​accident analysis model is a feedforward neural network that has performed multiple learning actions. The number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the dashcam. The feedforward neural network used includes a hidden layer, an input layer, and an output layer. The number of hidden layers is multiple and located between the input layer and the output layer. The number of hidden layers in the feedforward neural network used is inversely correlated with the value of the set frame rate. Thus, AI accident analysis models with different customized structures are designed for different dashcams, ensuring the reliability and stability of the intelligent analysis results of whether a vehicle turning accident will occur in the future time interval. Substantive Feature Four: To enable intelligent analysis of whether a vehicle turning accident will occur within a future time interval starting at a set time, multiple basic data sets are introduced. These basic data sets include the dashcam's resolution, the set frame rate, the duration of the time interval, the number of frames of the multiple past driving footage corresponding to the current time, and multiple first calculation values ​​and multiple second calculation values ​​corresponding to the multiple past driving footage corresponding to the current time. The comprehensive screening of these multiple basic data sets further ensures the reliability and stability of the intelligent analysis results of whether a vehicle turning accident will occur within the future time interval. Substantive Feature Five: Specifically, the road imaging sub-image in each frame of the past driving footage is uniformly divided into equal-area blocks to obtain uniformly divided blocks of the road imaging sub-image. The number of uniformly divided blocks in the road imaging sub-image with component standard deviations exceeding a limit is calculated as a first calculation value corresponding to the frame of the past driving footage. The number of curves in the road imaging sub-image in each frame of the driving footage with a curvature greater than or equal to a preset curvature value is calculated as a second calculation value corresponding to each frame of the driving footage, thereby completing the data structure design of two different calculation values ​​for intelligent analysis. Substantive feature six: In each learning action performed on the feedforward neural network, the accident occurrence identifier of a known vehicle turning accident within a certain past time interval is used as the single output content of the feedforward neural network, and the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the starting time of the certain past time interval, the multiple first calculation values ​​and the multiple second calculation values ​​corresponding to the multiple frames of past driving images corresponding to the starting time of the certain past time interval are used as multiple input contents of the feedforward neural network to complete this learning action, thereby ensuring the learning effect of each learning action of the feedforward neural network. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The embodiments of the present invention will be described below with reference to the accompanying drawings, in which: Figure 1 Schematic diagram of the working scenario of the dashcam data security storage and tracing system integrating blockchain technology according to the present invention.

[0011] Figure 2 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the first embodiment of the present invention.

[0012] Figure 3 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the second embodiment of the present invention.

[0013] Figure 4 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the third embodiment of the present invention.

[0014] Figure 5 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the fourth embodiment of the present invention.

[0015] Figure 6 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the fifth embodiment of the present invention.

[0016] Figure 7 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the sixth embodiment of the present invention. DETAILED DESCRIPTION

[0017] like Figure 1 The figure shows a schematic diagram of the working scenario of the dashcam data security storage and tracing system integrated with blockchain technology according to the present invention. The internet platform of the present invention more specifically relates to an accounting system for registering or indicating the operation of a machine.

[0018] The specific technical process of the present invention is as follows: Technical Process A: For each vehicle, each frame of driving footage recorded by its driving recorder is encrypted and stored on the blockchain. The driving recorder uses a set frame rate and has a corresponding resolution. Specifically, each frame of the driving picture is assigned a respective blockchain storage address in a blockchain distributed ledger storage mode for performing blockchain encrypted storage of the frame of the driving picture, and a consensus mechanism is used to perform blockchain encrypted storage; In this way, a data security storage mechanism can be provided for the subsequent tracing of multiple frames of driving images corresponding to the set time, ensuring the security and reliability of the called data; Technical process B: In order to realize intelligent analysis of whether a vehicle turning accident will occur in the future time interval starting at a set time, a customized AI accident analysis model is introduced, such as Figure 1 As shown; Specifically, the customized structural design of the AI ​​accident analysis model is mainly reflected in the following aspects: First: The AI ​​accident analysis model is a feedforward neural network that has performed multiple learning actions, and the number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the dashcam; Second: The feedforward neural network used in the AI ​​accident analysis model includes a hidden layer, an input layer, and an output layer. The number of hidden layers is multiple and is located between the input layer and the output layer. The number of hidden layers in the feedforward neural network used is inversely correlated with the value of the set frame rate. Third: In each learning action performed on the feedforward neural network, an accident occurrence identifier of a known vehicle turning accident within a certain past time interval is used as a single output content of the feedforward neural network, and the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the starting time of the certain past time interval, and the multiple first calculation values ​​and the multiple second calculation values ​​corresponding to the multiple frames of past driving images corresponding to the starting time of the certain past time interval are used as multiple input contents of the feedforward neural network to complete this learning action, thereby ensuring the learning effect of each learning action of the feedforward neural network; In this way, AI accident analysis models with different customized structures can be designed for different driving recorders, ensuring the reliability and stability of the intelligent analysis results of whether vehicle turning accidents will occur in the future time interval; Technical Process C: To achieve intelligent analysis of whether a vehicle turning accident will occur within a future time interval starting at a set time, multiple basic data points are introduced; Specifically, the multiple basic data include the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, and the multiple first accounting values ​​and the multiple second accounting values ​​corresponding to the multiple frames of past driving images corresponding to the current moment; like Figure 1 As shown, in addition to the multiple first calculation values ​​and the multiple second calculation values, the multiple basic data also include other auxiliary data for intelligent analysis, and the other auxiliary data include the resolution of the driving recorder, the set frame rate, the duration of the time interval, and the number of frames of the multiple past driving images corresponding to the current moment; More specifically, the road surface imaging sub-image in each frame of the past driving footage is uniformly divided into equal-area blocks to obtain uniformly divided blocks of the road surface imaging sub-image. The number of uniformly divided blocks in the road surface imaging sub-image with component standard deviations exceeding a limit is calculated as a first calculation value corresponding to the frame of the past driving footage. The number of curves in the road surface imaging sub-image in each frame of the driving footage with a curvature greater than or equal to a preset curvature value is calculated as a second calculation value corresponding to each frame of the driving footage, thereby completing the data structure design for two different calculation values ​​for intelligent analysis. In this way, through the comprehensive screening of the above-mentioned multiple basic data, the reliability and stability of the intelligent analysis results of whether vehicle turning accidents will occur in the future time interval are further guaranteed; Technical Process D: Use Technical Process B to customize the AI ​​accident analysis model designed for the current vehicle's dashcam. Based on the multiple basic data fully and comprehensively screened by Technical Process C, intelligently analyze whether the current vehicle will be involved in a turning accident within a future time period. For example, if the above intelligent analysis is performed at the current moment, the current time interval with the current moment as the starting moment belongs to a future time interval; Specifically, the intelligent analysis obtains an accident occurrence flag indicating whether the current vehicle will have a vehicle turning accident within a future time interval; Technical process E: Based on the intelligent analysis results of various accounting data collected according to the changing road conditions in technical process D, it indicates whether the current vehicle needs to enter the cautious operation mode, such as Figure 1 As shown; For example, when the intelligent analysis determines that a vehicle turning accident will occur within a future time interval starting at a set time, the vehicle is instructed to enter a cautious operation mode; otherwise, the vehicle is instructed to exit the cautious operation mode; In this way, an accounting system that indicates the operation of the machine is built, which improves the predictability and safety of driving for each vehicle equipped with a driving recorder; It can be seen that the present invention can perform blockchain encryption storage on each frame of driving images recorded by the driving recorder during the driving process of the vehicle, wherein each frame of the driving image is assigned a respective blockchain storage address in the blockchain distributed ledger storage mode for performing blockchain encryption storage on the frame of the driving image, and adopting a consensus mechanism to perform blockchain encryption storage, thereby providing a data security storage mechanism for the subsequent data tracing of multiple frames of driving images corresponding to a set time. More importantly, based on two different accounting values ​​obtained by analyzing various visual data of multiple frames of driving images before the set time, an intelligent analysis is performed on whether a vehicle turning accident will occur in a future time interval starting at the set time, and when the intelligent analysis determines that a vehicle turning accident will occur in a future time interval starting at the set time, the vehicle is instructed to enter a cautious operation mode; otherwise, the vehicle is instructed to exit the cautious operation mode, thereby building an accounting system for indicating the operation of the machine, thereby improving the predictability and safety of driving; The key points of this invention are: targeted collection of various accounting data based on changing road conditions, design of AI accident analysis models with different customized structures for different driving recorders, comprehensive screening of multiple basic data for intelligent analysis, and blockchain encrypted storage of road condition data combined with blockchain distributed ledger storage mode and consensus mechanism.

[0019] Below, a driving recorder data security storage and tracing system that integrates blockchain technology and is implemented by a smart chip of the present invention will be specifically described in the form of an embodiment.

[0020] First embodiment Figure 2 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the first embodiment of the present invention.

[0021] like Figure 2 As shown, the dashcam data security storage and tracing system integrating blockchain technology includes the following components: An encrypted storage device is used to encrypt and store each frame of driving footage recorded by the driving recorder during the vehicle's driving process in a blockchain. The driving recorder uses a set frame rate; For example, the dashcam installed in each vehicle has a set frame rate and a set resolution. For example, the dashcam has a set frame rate of 120 frames per second and a set resolution of 4096×2160, i.e., 4K ultra-high-definition resolution, which has a horizontal resolution of 4096 and a vertical resolution of 2160, and a total number of pixels of 4096×2160. A data tracing device is connected to the encrypted storage device and is used to trace back multiple frames of driving images before the current moment as multiple frames of past driving images corresponding to the current moment; In this way, through blockchain-based encrypted storage, the multiple frames of driving footage before the current moment are guaranteed to be safe, reliable and real data; a first calculation device, connected to the data tracing device, configured to uniformly divide the road surface imaging sub-image in each frame of the past driving image into equal areas to obtain uniformly divided blocks of the road surface imaging sub-image in the frame, and calculate the number of uniformly divided blocks in the road surface imaging sub-image in the frame whose component standard deviation exceeds a limit as a first calculation value corresponding to the frame of the past driving image; For example, when the road imaging sub-frames in each frame of the passing vehicle picture are evenly divided into equal areas, the even division of the equal areas is completed based on the pixel point accuracy; Here, the first calculated value corresponding to each frame of the past driving image represents the complexity of the road conditions in that frame of the past driving image. The multiple first calculated values ​​corresponding to multiple frames of driving images before the current moment represent the changes in the complexity of the road conditions before the current moment, and can serve as important basic data for analyzing whether a turning accident will occur in the subsequent time interval. A second calculation device, connected to the first calculation device, is used to calculate the number of curves in the road imaging sub-frame of each frame of the driving picture whose curvature is greater than or equal to a preset curvature value as a second calculation value corresponding to each frame of the driving picture; Here, the multiple second calculated values ​​corresponding to the multiple frames of driving images before the current moment represent the changes in the road surface conditions before the current moment, and can serve as another important basic data for analyzing whether a turning accident will occur in the subsequent time interval. an intelligent analysis device, connected to the first calculation device and the second calculation device, respectively, for using an AI accident analysis model to intelligently analyze an accident occurrence indicator of a vehicle turning accident within a current time interval based on the resolution of the driving recorder, a set frame rate, a duration of a time interval, a number of frames of a plurality of past driving images corresponding to a current moment, a plurality of first calculation values ​​corresponding to the plurality of past driving images corresponding to the current moment, and a plurality of second calculation values; Obviously, in addition to the first and second calculated values ​​reflecting changes in road conditions, the dashcam's resolution, set frame rate, duration of the time interval, and the number of frames of the current time frame corresponding to the previous driving footage are also used as auxiliary data to help complete the intelligent analysis of whether a turning accident will occur in the subsequent time interval. The AI ​​accident analysis model is a feedforward neural network that has performed multiple learning actions, and the number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the driving recorder. The feedforward neural network used includes a hidden layer, an input layer, and an output layer, and the number of the hidden layers is multiple and is located between the input layer and the output layer. For example, the number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the dashcam, including: when the resolution of the dashcam is 8K ultra-high-definition resolution, the number of learning actions performed by the feedforward neural network is 1000 times; when the resolution of the dashcam is 4K ultra-high-definition resolution, the number of learning actions performed by the feedforward neural network is 900 times; when the resolution of the dashcam is 2K ultra-high-definition resolution, the number of learning actions performed by the feedforward neural network is 800 times; when the resolution of the dashcam is high-definition resolution, the number of learning actions performed by the feedforward neural network is 700 times, and so on; In each learning action performed on the feedforward neural network, an accident occurrence identifier of a known vehicle turning accident within a certain past time interval is used as a single output content of the feedforward neural network, and the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the starting time of the certain past time interval, and the multiple first calculation values ​​and the multiple second calculation values ​​corresponding to the multiple frames of past driving images corresponding to the starting time of the certain past time interval are used as multiple input contents of the feedforward neural network to complete this learning action; Specifically, a numerical simulation mode can be selected to complete the test and simulation of each learning action performed by the feedforward neural network; The method comprises: performing uniform division of the road imaging sub-picture in each frame of the past driving image to obtain uniformly divided blocks of the road imaging sub-picture in the frame, and calculating the number of uniformly divided blocks in the road imaging sub-picture in the frame having component standard deviations exceeding a limit as a first calculation value corresponding to the frame of the past driving image. The method comprises: when a Y component standard deviation, a U component standard deviation, and / or a V component standard deviation of a certain uniformly divided block in the frame of the road imaging sub-picture exceeds a limit in a YUV space, using the certain uniformly divided block as the uniformly divided block in the frame of the road imaging sub-picture in which the component standard deviation exceeds a limit; For example, if the component values ​​of each pixel point of the provided past driving image are R component values, G component values, and B component values ​​in the RGB space, the Y component values, U component values, and V component values ​​of each pixel point of the past driving image in the YUV space can be obtained by using the RGB to YUV conversion formula; Specifically, the core formula for RGB to YUV conversion varies in different standards, but the commonly used formula under the ITU-R BT.601 standard is: Y = 0.299R + 0.587G + 0.114B; U = -0.1687R - 0.3313G + 0.5B + ​​128; V = 0.5R - 0.4187G - 0.0813B + 128; And wherein, when the Y component standard deviation, U component standard deviation and / or V component standard deviation of a certain uniformly divided block in the frame road imaging sub-image in the YUV space exceed the limit, the uniformly divided block is used as the uniformly divided block with the component standard deviation exceeding the limit in the frame road imaging sub-image, including: obtaining each Y component value / each U component value / each V component value corresponding to each component pixel point of the certain uniformly divided block, and when the standard deviation of each Y component value / each U component value / each V component value is greater than a preset standard deviation threshold, determining that the Y component standard deviation / U component value / V component value of the certain uniformly divided block in the YUV space exceeds the limit.

[0022] Second embodiment Figure 3 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the second embodiment of the present invention.

[0023] like Figure 3 As shown, compared with Figure 2 The dashcam data security storage and tracing system integrating blockchain technology also includes: The mode switching device is connected with the intelligent analysis device, and is used for instructing the vehicle to enter the cautious running mode when the accident occurrence identification of the vehicle turning accident in the current time interval obtained by the intelligent analysis indicates that the vehicle turning accident will occur in the current time interval. For example, the mode switching device can be implemented by using a programmable logic device, and is used for instructing the vehicle to enter the cautious running mode when the accident occurrence identification of the vehicle turning accident in the current time interval obtained by the intelligent analysis indicates that the vehicle turning accident will occur in the current time interval. For further example, the mode switching device can be implemented by using a programmable logic device, and is used for instructing the vehicle to enter the cautious running mode when the accident occurrence identification of the vehicle turning accident in the current time interval obtained by the intelligent analysis indicates that the vehicle turning accident will occur in the current time interval, and the programmable logic device is a FPGA chip designed in VHDL language. The mode switching device is further used for instructing the vehicle to exit the cautious running mode when the accident occurrence identification of the vehicle turning accident in the current time interval obtained by the intelligent analysis indicates that the vehicle turning accident will not occur in the current time interval.

[0024] Third embodiment Figure 4 An internal structure diagram of the driving recorder data security storage and tracing system based on the fusion of blockchain technology according to the third embodiment of the present application is shown.

[0025] As shown in Figure 4 Compared with Figure 3 , the driving recorder data security storage and tracing system based on the fusion of blockchain technology further comprises: The multi-layer learning device is connected with the intelligent analysis device, and is used for performing multiple learning actions on the feedforward neural network to obtain the feedforward neural network after the multiple learning actions are performed, and outputting the feedforward neural network after the multiple learning actions are performed as the AI accident analysis model. For example, the multi-layer learning device can be implemented by using a SOC chip, and is used for performing multiple learning actions on the feedforward neural network to obtain the feedforward neural network after the multiple learning actions are performed, and outputting the feedforward neural network after the multiple learning actions are performed as the AI accident analysis model. The multiple learning actions are performed on the feedforward neural network to obtain the feedforward neural network after the multiple learning actions are performed, and the feedforward neural network after the multiple learning actions are performed is outputted as the AI accident analysis model, which includes that each model parameter of the AI accident analysis model is used to complete the model representation of the AI accident analysis model.

[0026] Fourth embodiment Figure 5This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the fourth embodiment of the present invention.

[0027] like Figure 5 As shown, compared with Figure 2 The dashcam data security storage and tracing system integrating blockchain technology also includes: An instant display device, connected to the intelligent analysis device, for receiving an accident occurrence identifier of a vehicle turning accident within a current time interval and performing an instant display operation on the accident occurrence identifier of the vehicle turning accident within the current time interval; For example, the instant display device is connected to the intelligent analysis device, and is used to receive the accident occurrence identification of the vehicle turning accident within the current time interval, and perform an instant display operation on the accident occurrence identification of the vehicle turning accident within the current time interval, including: the instant display device is a liquid crystal display screen or an LCD display matrix.

[0028] Fifth embodiment Figure 6 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the fifth embodiment of the present invention.

[0029] like Figure 6 As shown, compared with Figure 2 The dashcam data security storage and tracing system integrating blockchain technology also includes: a Bluetooth transmission device connected to the intelligent analysis device, configured to receive an accident occurrence identifier of a vehicle turning accident within a current time interval, and wirelessly transmit the accident occurrence identifier of the vehicle turning accident within the current time interval to a Bluetooth headset of a vehicle driver via a Bluetooth transmission link; For example, a Bluetooth transmission device is connected to the intelligent analysis device, and is used to receive an accident occurrence identifier of a vehicle turning accident within a current time interval, and wirelessly transmit the accident occurrence identifier of the vehicle turning accident within the current time interval to the vehicle driver's Bluetooth headset through a Bluetooth transmission link, including: an accident occurrence identifier of a vehicle turning accident within the current time interval that can use a 0B10 value, indicating that a vehicle turning accident will not occur within the current time interval, and an accident occurrence identifier of a vehicle turning accident within the current time interval that can use a 0B11 value, indicating that a vehicle turning accident will occur within the current time interval.

[0030] Next, various embodiments of the present invention will be further described.

[0031] In each of the above embodiments, optionally, in the dashcam data security storage and tracing system integrating blockchain technology: An AI accident analysis model is used to intelligently analyze the accident occurrence identifier of a vehicle turning accident in the current time interval based on the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first accounting values ​​corresponding to the multiple frames of past driving images corresponding to the current moment, and the multiple second accounting values. The accident occurrence identifier includes: the current time interval takes the current moment as the starting moment, and the accident occurrence identifiers of different values ​​indicate whether a vehicle turning accident will occur in the current time interval; For example, the current time interval uses the current time as the starting time, and different values ​​of the accident occurrence flag indicate whether a vehicle turning accident will occur in the current time interval, including: the current time is 10:00 a.m., the current time interval is from 10:00 a.m. to 10:05 a.m., and the duration of each time interval is 5 minutes; wherein obtaining respective Y component values / U component values / V component values ​​corresponding to respective component pixels of the uniformly divided block, and when a standard deviation of the respective Y component values / U component values / V component values ​​is greater than a preset standard deviation threshold, determining that the Y component standard deviation / U component value / V component value of the uniformly divided block in the YUV space exceeds a limit includes: a value range of the Y component value / U component value / V component value of each component pixel of each uniformly divided block in the YUV space is between 0 and 255; The method further comprises: performing a road surface recognition operation on each frame of the past driving image to obtain a road surface imaging sub-image in each frame of the past driving image; and uniformly dividing the road surface imaging sub-image in each frame of the past driving image into equal-area uniformly divided blocks to obtain uniformly divided blocks of the road surface imaging sub-image in the frame of the past driving image, and calculating the number of uniformly divided blocks in the road surface imaging sub-image in the frame of the past driving image having a component standard deviation exceeding a limit as a first calculation value corresponding to the frame of the past driving image. And wherein, performing a road surface recognition operation on each frame of the past driving picture to obtain a road surface imaging sub-picture in each frame of the past driving picture includes: identifying the road surface component pixel points in each frame of the past driving picture based on the color imaging characteristics corresponding to the road surface, and fitting the various road surface component pixel points in the frame of the past driving picture to obtain the road surface imaging sub-picture in the frame of the past driving picture.

[0032] And in each of the above embodiments, optionally, in the dashcam data security storage and tracing system integrating blockchain technology: Each frame of driving footage recorded by the driving recorder during driving is encrypted and stored on the blockchain, wherein the driving recorder uses a set frame rate, including: allocating each frame of driving footage to a respective blockchain storage address in a blockchain distributed ledger storage mode for performing blockchain encrypted storage on the frame of driving footage, wherein the blockchain encrypted storage is performed using a consensus mechanism; The so-called blockchain distributed ledger storage mode is to scatter and distribute the originally centrally stored data to different blockchain storage nodes for storage, thereby eliminating the possibility of a single power center interfering with the entire storage system. The so-called consensus mechanism is an important component of blockchain technology. Blockchain, as a data structure for storing data in chronological order, can support different consensus mechanisms. The goal of the blockchain consensus mechanism is to make all honest nodes save a consistent blockchain view while satisfying two properties: 1) consistency, that is, the prefix part of the blockchain saved by all honest nodes is completely the same; 2) validity, that is, the information published by a certain honest node will eventually be recorded in the blockchain of all other honest nodes. Among them, the blockchain encryption storage is performed on each frame of driving picture recorded by the driving recorder during the driving of the vehicle, and the driving recorder adopting the set frame rate further comprises: the uniform interval distribution of each record timestamp corresponding to each frame of driving picture recorded by the driving recorder during the driving of the vehicle on the time axis. Among them, the resolution of the driving recorder, the set frame rate, the duration of the time interval, the frame number of the plurality of past driving pictures corresponding to the current time, and the plurality of first and second calculation values corresponding to the plurality of past driving pictures corresponding to the current time are synchronously input into the AI accident analysis model, and the AI accident analysis model is run to obtain the accident occurrence identifier of the vehicle turning accident in the current time interval output by the AI accident analysis model. Among them, the resolution of the driving recorder, the set frame rate, the duration of the time interval, the frame number of the plurality of past driving pictures corresponding to the current time, and the plurality of first and second calculation values corresponding to the plurality of past driving pictures corresponding to the current time are synchronously input into the AI accident analysis model, and the AI accident analysis model is run to obtain the accident occurrence identifier of the vehicle turning accident in the current time interval output by the AI accident analysis model. For example, the accident occurrence identifier of the vehicle turning accident in the current time interval, the resolution of the driving recorder, the set frame rate, the duration of the time interval, the frame number of the plurality of past driving pictures corresponding to the current time, and the plurality of first and second calculation values corresponding to the plurality of past driving pictures corresponding to the current time are all in the numerical value representation form after numerical value normalization processing. And wherein, the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first calculation values ​​and the multiple second calculation values ​​corresponding to the multiple frames of past driving images corresponding to the current moment are synchronously input into the AI ​​accident analysis model, and the AI ​​accident analysis model is run to obtain the accident occurrence mark of the vehicle turning accident in the current time interval output by the AI ​​accident analysis model, which also includes: the resolution of the driving recorder is the vertical resolution of the driving recorder and the horizontal resolution of the driving recorder.

[0033] Sixth embodiment Figure 7 This is a diagram showing the internal structure of a dashcam data security storage and tracing system integrating blockchain technology according to the sixth embodiment of the present invention.

[0034] like Figure 7 As shown, the system for automatically generating and publishing all-media content includes a memory and multiple processors, wherein the memory stores a computer program, and the computer program is configured to be executed by the multiple processors to complete the following steps: Step 71: Each frame of driving footage recorded by the driving recorder during the driving process is encrypted and stored in the blockchain, and the driving recorder uses a set frame rate; For example, the dashcam installed in each vehicle has a set frame rate and a set resolution. For example, the dashcam has a set frame rate of 120 frames per second and a set resolution of 4096×2160, i.e., 4K ultra-high-definition resolution, which has a horizontal resolution of 4096 and a vertical resolution of 2160, and a total number of pixels of 4096×2160. Step 72: tracing back multiple frames of driving images before the current moment as multiple frames of past driving images corresponding to the current moment; In this way, through blockchain-based encrypted storage, the multiple frames of driving footage before the current moment are guaranteed to be safe, reliable and authentic data; Step 73: Evenly divide the road imaging sub-image in each frame of the past driving image into equal-area blocks to obtain evenly divided blocks of the road imaging sub-image. Calculate the number of evenly divided blocks in the road imaging sub-image with component standard deviations exceeding a limit as a first calculated value corresponding to the frame of the past driving image. For example, when the road imaging sub-frames in each frame of the passing vehicle picture are evenly divided into equal areas, the even division of the equal areas is completed based on the pixel point accuracy; Here, the first calculated value corresponding to each frame of the past driving image represents the complexity of the road conditions in that frame of the past driving image. The multiple first calculated values ​​corresponding to multiple frames of driving images before the current moment represent the changes in the complexity of the road conditions before the current moment, and can serve as important basic data for analyzing whether a turning accident will occur in the subsequent time interval. Step 74: Calculate the number of curves in each frame of the road imaging sub-frame whose curvature is greater than or equal to a preset curvature value as a second calculation value corresponding to each frame of the driving image; Here, the multiple second calculated values ​​corresponding to the multiple frames of driving images before the current moment represent the changes in the road surface conditions before the current moment, and can serve as another important basic data for analyzing whether a turning accident will occur in the subsequent time interval. Step 75: Using the AI ​​accident analysis model, intelligently analyze the accident occurrence indicator of the vehicle turning accident in the current time interval based on the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving footage corresponding to the current moment, the multiple first calculation values ​​corresponding to the multiple frames of past driving footage corresponding to the current moment, and the multiple second calculation values; Obviously, in addition to the first and second calculated values ​​reflecting changes in road conditions, the dashcam's resolution, set frame rate, duration of the time interval, and the number of frames of the current time frame corresponding to the previous driving footage are also used as auxiliary data to help complete the intelligent analysis of whether a turning accident will occur in the subsequent time interval. The AI ​​accident analysis model is a feedforward neural network that has performed multiple learning actions, and the number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the driving recorder. The feedforward neural network used includes a hidden layer, an input layer, and an output layer, and the number of the hidden layers is multiple and is located between the input layer and the output layer. For example, the number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the dashcam, including: when the resolution of the dashcam is 8K ultra-high-definition resolution, the number of learning actions performed by the feedforward neural network is 1000 times; when the resolution of the dashcam is 4K ultra-high-definition resolution, the number of learning actions performed by the feedforward neural network is 900 times; when the resolution of the dashcam is 2K ultra-high-definition resolution, the number of learning actions performed by the feedforward neural network is 800 times; when the resolution of the dashcam is high-definition resolution, the number of learning actions performed by the feedforward neural network is 700 times, and so on; In each learning action performed on the feedforward neural network, an accident occurrence identifier of a known vehicle turning accident within a certain past time interval is used as a single output content of the feedforward neural network, and the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the starting time of the certain past time interval, and the multiple first calculation values ​​and the multiple second calculation values ​​corresponding to the multiple frames of past driving images corresponding to the starting time of the certain past time interval are used as multiple input contents of the feedforward neural network to complete this learning action; Specifically, a numerical simulation mode can be selected to complete the test and simulation of each learning action performed by the feedforward neural network; The method comprises: performing uniform division of the road imaging sub-picture in each frame of the past driving image to obtain uniformly divided blocks of the road imaging sub-picture in the frame, and calculating the number of uniformly divided blocks in the road imaging sub-picture in the frame having component standard deviations exceeding a limit as a first calculation value corresponding to the frame of the past driving image. The method comprises: when a Y component standard deviation, a U component standard deviation, and / or a V component standard deviation of a certain uniformly divided block in the frame of the road imaging sub-picture exceeds a limit in a YUV space, using the certain uniformly divided block as the uniformly divided block in the frame of the road imaging sub-picture in which the component standard deviation exceeds a limit; For example, if the component values ​​of each pixel point of the provided past driving image are R component values, G component values, and B component values ​​in the RGB space, the Y component values, U component values, and V component values ​​of each pixel point of the past driving image in the YUV space can be obtained by using the RGB to YUV conversion formula; Specifically, the core formula for RGB to YUV conversion varies in different standards, but the commonly used formula under the ITU-R BT.601 standard is: Y = 0.299R + 0.587G + 0.114B; U = -0.1687R - 0.3313G + 0.5B + ​​128; V = 0.5R - 0.4187G - 0.0813B + 128; And wherein, when the Y component standard deviation, U component standard deviation and / or V component standard deviation of a certain uniformly divided block in the frame road imaging sub-image in the YUV space exceed the limit, the uniformly divided block is used as the uniformly divided block with the component standard deviation exceeding the limit in the frame road imaging sub-image, including: obtaining each Y component value / each U component value / each V component value corresponding to each component pixel point of the certain uniformly divided block, and when the standard deviation of each Y component value / each U component value / each V component value is greater than a preset standard deviation threshold, determining that the Y component standard deviation / U component value / V component value of the certain uniformly divided block in the YUV space exceeds the limit.

[0035] In addition, in the dashcam data security storage and tracing system integrating blockchain technology according to the present invention: The AI ​​accident analysis model is a feedforward neural network that has performed multiple learning actions, and the number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the driving recorder. The feedforward neural network used includes a hidden layer, an input layer, and an output layer. The number of hidden layers is multiple and is located between the input layer and the output layer. The number of hidden layers in the feedforward neural network used is inversely correlated with the value of the set frame rate. The number of hidden layers in the feedforward neural network used and the value of the set frame rate are inversely correlated, including: using a numerical mapping formula to express a numerical mapping relationship in which the number of hidden layers in the feedforward neural network used and the value of the set frame rate are inversely correlated; For example, a MATLAB toolbox may be selected to complete the test and simulation of a numerical mapping process that uses a numerical mapping formula to represent a numerical mapping relationship inversely related to the number of hidden layers in a feedforward neural network and a set frame rate value. And wherein, in the numerical mapping formula, the value of the set frame rate is the input parameter of the numerical mapping formula, and the number of hidden layers in the feedforward neural network used that is inversely associated with the value of the set frame rate is the output parameter of the numerical mapping formula.

[0036] The foregoing description of the exemplary embodiments of the present invention has been provided for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Obviously, many modifications and variations will be apparent to those skilled in the art. The exemplary embodiments were chosen and described in order to best illustrate the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the various embodiments and variations of the invention as they may be adapted for the specific application contemplated. It is intended that the scope of the invention be defined by the appended claims and their equivalents.

Claims

1. A driving recorder data security storage and tracing system integrating blockchain technology, characterized by: The system comprises: An encrypted storage device is used to encrypt and store each frame of driving footage recorded by the driving recorder during the vehicle's driving process in a blockchain. The driving recorder uses a set frame rate; A data tracing device is connected to the encrypted storage device and is used to trace back multiple frames of driving images before the current moment as multiple frames of past driving images corresponding to the current moment; a first calculation device, connected to the data tracing device, configured to uniformly divide the road surface imaging sub-image in each frame of the past driving image into equal areas to obtain uniformly divided blocks of the road surface imaging sub-image in the frame, and calculate the number of uniformly divided blocks in the road surface imaging sub-image in the frame whose component standard deviation exceeds a limit as a first calculation value corresponding to the frame of the past driving image; A second calculation device, connected to the first calculation device, is used to calculate the number of curves in the road imaging sub-frame of each frame of the driving picture whose curvature is greater than or equal to a preset curvature value as a second calculation value corresponding to each frame of the driving picture; The intelligent analysis device is connected to the first calculation device and the second calculation device respectively, and is used to use the AI ​​accident analysis model to intelligently analyze the accident occurrence signs of vehicle turning accidents in the current time interval according to the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving pictures corresponding to the current moment, the multiple first calculation values ​​corresponding to the multiple frames of past driving pictures corresponding to the current moment, and the multiple second calculation values.

2. The dashcam data security storage and tracing system integrating blockchain technology as claimed in claim 1 is characterized by: The AI ​​accident analysis model is a feedforward neural network that has performed multiple learning actions, and the number of learning actions performed by the feedforward neural network is positively correlated with the resolution of the driving recorder. The feedforward neural network used includes a hidden layer, an input layer, and an output layer, and the number of the hidden layers is multiple and is located between the input layer and the output layer. Among them, in each learning action performed on the feedforward neural network, the accident occurrence identifier of a known vehicle turning accident within a certain past time interval is used as a single output content of the feedforward neural network, and the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the starting time of the certain past time interval, the multiple first calculation values ​​and the multiple second calculation values ​​corresponding to the multiple frames of past driving images corresponding to the starting time of the certain past time interval are used as multiple input contents of the feedforward neural network to complete this learning action.

3. The dashcam data security storage and tracing system integrating blockchain technology as claimed in claim 2 is characterized by: Performing uniform division of the road imaging sub-picture in each frame of the past driving picture to obtain uniformly divided blocks of the road imaging sub-picture in the frame, and calculating the number of uniformly divided blocks in the road imaging sub-picture in the frame with component standard deviations exceeding a limit as a first calculation value corresponding to the frame of the past driving picture includes: when a Y component standard deviation, a U component standard deviation, and / or a V component standard deviation of a certain uniformly divided block in the frame of the road imaging sub-picture exceeds a limit in the YUV space, using the certain uniformly divided block as the uniformly divided block in the frame of the road imaging sub-picture with component standard deviation exceeding a limit; When the Y component standard deviation, U component standard deviation, and / or V component standard deviation of a certain uniformly divided block in the frame of road surface imaging sub-image exceed a limit in the YUV space, determining the certain uniformly divided block as the uniformly divided block with the component standard deviation exceeding the limit in the frame of road surface imaging sub-image includes: obtaining respective Y component values / U component values / V component values ​​corresponding to respective component pixel points of the certain uniformly divided block, and when the standard deviation of the respective Y component values / U component values / V component values ​​is greater than a preset standard deviation threshold, determining that the Y component standard deviation / U component value / V component value of the certain uniformly divided block in the YUV space exceeds a limit.

4. A driving recorder data security storage and tracing system integrating blockchain technology as claimed in claim 3, characterized in that: The system further comprises: a mode switching device connected to the intelligent analysis device, for instructing the vehicle to enter a cautious operation mode when an accident occurrence indicator of a vehicle turning accident within a current time interval obtained by the intelligent analysis indicates that a vehicle turning accident will occur within the current time interval; The mode switching device is further configured to instruct the vehicle to exit the cautious operation mode when the accident occurrence indicator of the vehicle turning accident obtained through intelligent analysis within the current time interval indicates that the vehicle turning accident will not occur within the current time interval.

5. The driving recorder data security storage and tracing system integrating blockchain technology as claimed in claim 3 is characterized in that: The system further comprises: A multi-layer learning device, connected to the intelligent analysis device, is used to perform multiple learning actions on the feedforward neural network to obtain a feedforward neural network after the multiple learning actions are performed, and output the feedforward neural network after the multiple learning actions as an AI accident analysis model; Among them, performing multiple learning actions on the feedforward neural network to obtain the feedforward neural network after performing the multiple learning actions, and outputting the feedforward neural network after performing the multiple learning actions as the AI ​​accident analysis model includes: using various model parameters of the AI ​​accident analysis model to complete the model representation of the AI ​​accident analysis model.

6. The driving recorder data security storage and tracing system integrating blockchain technology as claimed in claim 3 is characterized in that: The system further comprises: The instant display device is connected to the intelligent analysis device and is used to receive the accident occurrence identification of the vehicle turning accident in the current time interval and perform an instant display operation on the accident occurrence identification of the vehicle turning accident in the current time interval.

7. The driving recorder data security storage and tracing system integrating blockchain technology as claimed in claim 3 is characterized in that: The system further comprises: A Bluetooth transmission device is connected to the intelligent analysis device and is used to receive the accident occurrence identifier of the vehicle turning accident in the current time interval, and wirelessly transmit the accident occurrence identifier of the vehicle turning accident in the current time interval to the Bluetooth headset of the vehicle driver through the Bluetooth transmission link.

8. A dashcam data security storage and tracing system integrating blockchain technology as described in any one of claims 3-7, characterized in that: An AI accident analysis model is used to intelligently analyze the accident occurrence identifier of a vehicle turning accident in the current time interval based on the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first accounting values ​​corresponding to the multiple frames of past driving images corresponding to the current moment, and the multiple second accounting values. The accident occurrence identifier includes: the current time interval takes the current moment as the starting moment, and the accident occurrence identifiers of different values ​​indicate whether a vehicle turning accident will occur in the current time interval; wherein obtaining respective Y component values / U component values / V component values ​​corresponding to respective component pixels of the uniformly divided block, and when a standard deviation of the respective Y component values / U component values / V component values ​​is greater than a preset standard deviation threshold, determining that the Y component standard deviation / U component value / V component value of the uniformly divided block in the YUV space exceeds a limit includes: a value range of the Y component value / U component value / V component value of each component pixel of each uniformly divided block in the YUV space is between 0 and 255; The method further comprises: performing a road surface recognition operation on each frame of the past driving image to obtain a road surface imaging sub-image in each frame of the past driving image; and uniformly dividing the road surface imaging sub-image in each frame of the past driving image into equal-area uniformly divided blocks to obtain uniformly divided blocks of the road surface imaging sub-image in the frame of the past driving image, and calculating the number of uniformly divided blocks in the road surface imaging sub-image in the frame of the past driving image having a component standard deviation exceeding a limit as a first calculation value corresponding to the frame of the past driving image. Among them, performing a road surface recognition operation on each frame of the past driving picture to obtain a road surface imaging sub-picture in each frame of the past driving picture includes: identifying the road surface component pixel points in each frame of the past driving picture based on the color imaging characteristics corresponding to the road surface, and fitting the various road surface component pixel points in the frame of the past driving picture to obtain the road surface imaging sub-picture in the frame of the past driving picture.

9. A dashcam data security storage and tracing system integrating blockchain technology as described in any one of claims 3-7, characterized in that: Each frame of driving footage recorded by the driving recorder during driving is encrypted and stored on the blockchain, wherein the driving recorder uses a set frame rate, including: allocating each frame of driving footage to a respective blockchain storage address in a blockchain distributed ledger storage mode for performing blockchain encrypted storage on the frame of driving footage, wherein the blockchain encrypted storage is performed using a consensus mechanism; Among them, each frame of driving footage recorded by the driving recorder during the vehicle's driving process is encrypted and stored on the blockchain. The driving recorder uses a set frame rate and further includes: the recording timestamps corresponding to each frame of driving footage recorded by the driving recorder during the vehicle's driving process are evenly spaced on the time axis; The resolution of the dashcam, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving footage corresponding to the current moment, the multiple first calculation values ​​and the multiple second calculation values ​​corresponding to the multiple frames of past driving footage corresponding to the current moment are synchronously input into the AI ​​accident analysis model, and the AI ​​accident analysis model is run to obtain an accident occurrence identifier of a vehicle turning accident within the current time interval output by the AI ​​accident analysis model; Among them, the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first accounting values ​​and the multiple second accounting values ​​corresponding to the multiple frames of past driving images corresponding to the current moment are synchronously input into the AI ​​accident analysis model, and the AI ​​accident analysis model is run to obtain the accident occurrence identifier of the vehicle turning accident in the current time interval output by the AI ​​accident analysis model, including: the accident occurrence identifier of the vehicle turning accident in the current time interval, the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first accounting values ​​and the multiple second accounting values ​​corresponding to the multiple frames of past driving images corresponding to the current moment are all numerical representations after numerical normalization processing; Among them, the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first calculation values ​​corresponding to the multiple frames of past driving images corresponding to the current moment, and the multiple second calculation values ​​are synchronously input into the AI ​​accident analysis model, and the AI ​​accident analysis model is run to obtain the accident occurrence mark of the vehicle turning accident in the current time interval output by the AI ​​accident analysis model, which also includes: the resolution of the driving recorder is the vertical resolution of the driving recorder and the horizontal resolution of the driving recorder.

10. A dashcam data security storage and tracing system integrating blockchain technology, the system comprising a memory and multiple processors, the memory storing a computer program, characterized in that: The computer program is configured to be executed by the plurality of processors to perform the following steps: Each frame of driving footage recorded by the driving recorder during the vehicle's driving process is encrypted and stored on the blockchain. The driving recorder uses a set frame rate; At the current moment, multiple frames of driving images before the current moment are traced back to serve as multiple frames of past driving images corresponding to the current moment; Performing uniform segmentation of equal areas on the road imaging sub-image in each frame of the past driving image to obtain uniformly segmented blocks of the road imaging sub-image in the frame, and calculating the number of uniformly segmented blocks in the road imaging sub-image in the frame whose component standard deviation exceeds a limit as a first calculation value corresponding to the frame of the past driving image; Calculating the number of curves in the road imaging sub-frame of each frame of the driving picture whose curvature is greater than or equal to a preset curvature value as a second calculation value corresponding to each frame of the driving picture; An AI accident analysis model is used to intelligently analyze the accident occurrence signs of vehicle turning accidents in the current time interval based on the resolution of the driving recorder, the set frame rate, the duration of the time interval, the number of frames of the multiple frames of past driving images corresponding to the current moment, the multiple first calculation values ​​corresponding to the multiple frames of the past driving images corresponding to the current moment, and the multiple second calculation values.

Images