Loop recording method and device of vehicle-mounted intelligent interactive all-in-one machine

Abstract

The invention provides a cyclic recording method and device for a vehicle-mounted intelligent interaction all-in-one machine, and belongs to the field of data processing, and the method comprises the steps: obtaining a vehicle-mounted image of a target vehicle; dividing the vehicle-mounted image into a plurality of vehicle-mounted image segments, and storing the plurality of vehicle-mounted image segments in a storage device; determining whether a target vehicle-mounted image segment related to vehicle collision exists in the plurality of vehicle-mounted image segments; if the target vehicle-mounted image segment exists, the target vehicle-mounted image segment is subjected to locking marking, and the vehicle-mounted image segments which are not subjected to locking marking can be circularly covered in the storage device. And the storage integrity of the collision key image and the continuity of cyclic recording of the vehicle-mounted image can be considered.

Description

Technical Field

[0001] This application belongs to the field of data processing technology, specifically relating to a loop recording method and device for an in-vehicle intelligent interactive all-in-one machine. Background Technology

[0002] In the field of vehicle video recording, the loop recording function of in-vehicle intelligent interactive all-in-one machines has been widely used. However, the current technology for storing and managing vehicle images still has significant technical shortcomings. Current vehicle video recording solutions mostly use continuous storage or simple segmented storage, lacking specific identification and protection mechanisms for vehicle collision-related images. They cannot accurately distinguish between ordinary driving images and critical collision-related images, leading to the accidental deletion of important images after a collision during the loop overwriting process, making it difficult to retain effective accident evidence.

[0003] Furthermore, if all recorded images are indiscriminately saved without loop overwriting, the storage space of the device will quickly fill up, making it impossible for the device to continuously record vehicle images. This approach fails to meet the recording needs of vehicles during long-term driving and makes it difficult to effectively retain critical collision images under limited storage resources. It also makes it difficult to balance the continuity of loop recording of vehicle images with the integrity of the preserved critical collision images, causing many inconveniences to the practical application of vehicle image recording. Summary of the Invention

[0004] This application provides a method and device for loop recording of in-vehicle intelligent interactive all-in-one machine, so as to balance the integrity of the preservation of key collision images and the continuity of in-vehicle image loop recording.

[0005] This application provides a method for loop recording of an in-vehicle intelligent interactive all-in-one machine, including:

[0006] Acquire in-vehicle images of the target vehicle;

[0007] The vehicle image is divided into multiple vehicle image segments, and the multiple vehicle image segments are stored in a storage device;

[0008] Determine whether any of the plurality of vehicle image segments are related to a vehicle collision;

[0009] If the target vehicle image segment exists, the target vehicle image segment is locked. Vehicle image segments that are not locked are cyclically overwritten in the storage device.

[0010] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, the method further includes: determining the number of loop recordings of the storage device; if the number of loop recordings is greater than a preset number, determining a reference in-vehicle image segment from the target in-vehicle image segment that has been locked; editing the reference in-vehicle image segment to obtain an edited image segment; locking the edited image segment and deleting the target in-vehicle image segment that has been locked.

[0011] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, after determining the reference in-vehicle image segment from the target in-vehicle image segment with lock marking, the method further includes: deleting the lock markings of other target in-vehicle image segments besides the reference in-vehicle image segment from the target in-vehicle image segment with lock marking.

[0012] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, the step of editing the reference in-vehicle image segment to obtain the edited image segment includes: determining the collision time from the reference in-vehicle image segment; determining the collision time period with the collision time as the center; extracting a first collision image segment of the collision time period from the reference in-vehicle image segment; analyzing the first collision image segment to obtain the analysis result; and editing the first collision image segment according to the analysis result to obtain the edited image segment.

[0013] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, the step of analyzing the first collision image segment to obtain the analysis result includes: acquiring the target object included in each frame of the first collision image segment and the motion information of the target object; determining the change in the motion state of the target object based on the target object included in each frame of the image and the motion information of the target object; determining multiple target image frames based on the change in the motion state, the multiple target image frames indicating the complete collision process; and obtaining the analysis result based on the multiple target image frames.

[0014] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, determining whether there is a target in-vehicle image segment related to a vehicle collision among the plurality of in-vehicle image segments includes: acquiring gravity monitoring information from the gravity sensor of the target vehicle; determining target gravity monitoring information from the gravity monitoring information, wherein the target gravity monitoring information is the gravity detection information at the moment of collision of the target vehicle; determining the target time for acquiring the target gravity monitoring information; and determining the in-vehicle image segment that includes the target time among the plurality of in-vehicle image segments as the target in-vehicle image segment.

[0015] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, the step of determining the target gravity monitoring information from the gravity monitoring information includes: obtaining the three-axis acceleration information of the target vehicle based on the gravity monitoring information, wherein the three axes indicate two horizontal directions and one vertical direction; generating time-acceleration waveform data for each axis based on the three-axis acceleration information; obtaining the vehicle type of the target vehicle; obtaining the driving scene of the target vehicle in each time period; and determining the target gravity monitoring information based on the vehicle type, the driving scene, and the time-acceleration waveform data for each axis.

[0016] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, the step of determining the target gravity monitoring information based on the vehicle type, the driving scenario, and the time-acceleration waveform data of each axis includes: determining the waveform characteristics in each time period based on the time-acceleration waveform data of each axis, wherein the waveform characteristics include at least one of peak acceleration, waveform rise slope, or waveform duration; determining the waveform characteristic threshold of each axis in the three-axis acceleration information based on the vehicle type and the driving scenario; and determining the target gravity monitoring information based on the waveform characteristics in each time period and the waveform characteristic threshold of each axis.

[0017] According to the loop recording method of the in-vehicle intelligent interactive all-in-one machine provided in this application, the step of determining whether there is a target in-vehicle image segment related to vehicle collision among the plurality of in-vehicle image segments includes: obtaining an emergency locking command; determining the locking time of the emergency locking command; obtaining a target in-vehicle image segment according to the locking time, wherein the target in-vehicle image segment includes the in-vehicle image at the locking time.

[0018] This application also provides an in-vehicle intelligent interactive all-in-one machine, including an in-vehicle controller, the in-vehicle controller being used for:

[0019] Acquire in-vehicle images of the target vehicle;

[0020] The vehicle image is divided into multiple vehicle image segments, and the multiple vehicle image segments are stored in a storage device;

[0021] Determine whether any of the plurality of vehicle image segments are related to a vehicle collision;

[0022] If the target vehicle image segment exists, the target vehicle image segment is locked. Vehicle image segments that are not locked are cyclically overwritten in the storage device.

[0023] This application also provides a loop recording device for an in-vehicle intelligent interactive all-in-one machine, including:

[0024] The acquisition unit is used to acquire the in-vehicle image of the target vehicle;

[0025] A segmentation unit is used to divide the vehicle image into multiple vehicle image segments and store the multiple vehicle image segments in a storage device;

[0026] A determining unit is used to determine whether there is a target vehicle image segment related to a vehicle collision among the plurality of vehicle image segments;

[0027] The locking marking unit is used to lock the target vehicle image segment if it exists, wherein vehicle image segments that are not locked will be cyclically overwritten in the storage device.

[0028] This application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the loop recording method of the in-vehicle intelligent interactive all-in-one machine as described above.

[0029] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the loop recording method of any of the above-described in-vehicle intelligent interactive all-in-one machines.

[0030] According to the loop recording method and device for an in-vehicle intelligent interactive all-in-one machine provided in this application, the in-vehicle controller first acquires the in-vehicle image of the target vehicle, then divides the in-vehicle image into multiple in-vehicle image segments, and stores the multiple in-vehicle image segments in a storage device. Next, it is determined whether there is a target in-vehicle image segment related to a vehicle collision among the multiple in-vehicle image segments. Finally, if the target in-vehicle image segment exists, it is locked. Unlocked in-vehicle image segments are cyclically overwritten in the storage device. This solution, by first acquiring the in-vehicle image and dividing it into multiple segments for storage, then identifying and locking the image segments related to a vehicle collision, and simultaneously performing cyclic overwriting on the unlocked image segments, can effectively retain key image segments related to vehicle collisions and prevent accidental deletion, while rationally utilizing the storage space of the storage device to achieve continuous loop recording of in-vehicle images, thus balancing the integrity of the key collision images and the continuity of loop recording of in-vehicle images. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a flowchart illustrating the loop recording method for the in-vehicle intelligent interactive all-in-one machine provided in this application.

[0033] Figure 2 This is one of the schematic diagrams of the management interface for the vehicle-mounted video clips provided in this application.

[0034] Figure 3 This is the second schematic diagram of the management interface for the vehicle-mounted video clips provided in this application.

[0035] Figure 4 This is a schematic diagram of the emergency locking interface of the vehicle-mounted video footage provided in this application.

[0036] Figure 5 This is a structural schematic diagram of the in-vehicle intelligent interactive all-in-one machine provided in this application. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0039] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0040] Existing vehicle image storage management methods either lack a specific identification and protection mechanism for vehicle collision-related images, making it easy for critical collision images to be accidentally deleted during loop overwriting, or they indiscriminately save all recorded images without performing loop overwriting, quickly filling up storage space and preventing the device from recording continuously. It is difficult to balance the continuity of vehicle image loop recording and the integrity of critical collision images under the premise of limited storage resources, which brings many inconveniences to practical applications.

[0041] To address the aforementioned problems, this application provides a method and device for loop recording of an in-vehicle intelligent interactive all-in-one machine. The following detailed description, in conjunction with specific embodiments, further illustrates this application.

[0042] Please see Figure 1 , Figure 1 This is a flowchart illustrating the loop recording method for the in-vehicle intelligent interactive all-in-one machine provided in this application. The loop recording method for the in-vehicle intelligent interactive all-in-one machine includes the following steps.

[0043] S101, acquire the in-vehicle image of the target vehicle.

[0044] Among them, in-vehicle imaging refers to real-time image data in various scenarios during vehicle operation. Real-time images of the area in front of and behind the vehicle can be captured using front-view and rear-view cameras mounted on the target vehicle. Side-view and interior cameras can also capture images of the vehicle's sides and interior. Furthermore, a 360-degree panoramic camera can be combined to capture a panoramic view of the vehicle's surroundings. Various cameras can capture images at preset frame rates and resolutions, supporting the acquisition and transmission of multiple video formats. The resolution parameters can be flexibly adjusted based on the hardware performance of the in-vehicle intelligent interactive all-in-one machine and actual recording needs. The entire image acquisition process can be uniformly controlled by the in-vehicle controller of the intelligent interactive all-in-one machine. Image acquisition is automatically triggered after the vehicle starts and stops when the vehicle stops. Alternatively, it can continue acquiring images after the vehicle stops, or it can start acquiring images when a collision event is detected after the vehicle stops.

[0045] S102, the vehicle image is divided into multiple vehicle image segments, and the multiple vehicle image segments are stored in a storage device.

[0046] The vehicle-mounted video segments can be divided equally according to a preset time length. The division duration can be flexibly set according to the actual application needs of the vehicle-mounted intelligent interactive all-in-one machine, such as 1 minute, 3 minutes, or 5 minutes. Continuously acquired vehicle-mounted video is cut into independent and continuous video segments according to this time standard. Each video segment is numbered sequentially according to the recording time, ensuring the temporality and integrity of the video segments. Simultaneously, the segmentation process can be executed in real time by the vehicle-mounted controller, completing the segmentation synchronously during the continuous acquisition of vehicle-mounted video, eliminating the need to wait for the video acquisition to finish before batch processing, ensuring the synchronization of video segmentation and acquisition.

[0047] Storing in-vehicle video clips can be achieved using the storage device integrated into the in-vehicle intelligent interactive all-in-one machine. The segmented video clips can be stored in the storage device according to a preset file format. During storage, the video clips can be stored sequentially according to their recording time, assigning each video clip a unique storage address and file identifier for easy retrieval and location of the corresponding video clip later. Simultaneously, the storage device can continuously write to and store the segmented video clips. Through cyclic recording, when the storage space reaches a preset threshold, the currently stored video can be overwritten sequentially with previously stored videos based on their storage time. For example... Figure 2 As shown, users can view stored video clips on the display interface of the in-vehicle intelligent interactive all-in-one machine. This includes directly jumping to the stored in-vehicle video clips for the current day by clicking the "Date" control. S103, determine whether there is a target in-vehicle video clip related to a vehicle collision among the multiple in-vehicle video clips.

[0048] The target vehicle video clips include those that reflect the entire process of a vehicle collision. Specifically, these include video clips capturing the instant of the collision, as well as video clips that fully present the cause, process, and result of the collision within a preset time period before and after the collision. To determine the existence of a target vehicle video clip, it is first confirmed whether a collision event occurred with the target vehicle. Then, based on the time information corresponding to the collision event, video clips containing that time information are matched from the ordered stored vehicle video clips and identified as target vehicle video clips related to the vehicle collision. Furthermore, during the determination process, the recording time and related information of each vehicle video clip are comprehensively verified to ensure that the selected target vehicle video clips accurately correspond to the vehicle collision event.

[0049] S104, if the target vehicle image segment exists, then the target vehicle image segment is locked.

[0050] In this process, vehicle video clips without lock tags are cyclically overwritten in the storage device. Cyclic recording involves a complete storage management process, including sequential writing of vehicle video clips to the storage device, space monitoring, and automatic overwriting. Specifically, vehicle video clips are first written to the storage device sequentially according to their recording time. Simultaneously, the remaining storage space on the storage device is monitored in real-time by the vehicle controller of the in-vehicle intelligent interactive all-in-one machine. When the remaining storage space reaches a preset threshold, a cyclic overwriting operation is triggered. Following a "first-in, first-out" principle, the earliest recorded and unlocked vehicle video clips are deleted segment by segment to free up storage space for newly acquired and segmented vehicle video clips. New video clips are then immediately written to the free space in the storage device, thus forming an uninterrupted cyclic recording process.

[0051] When locking a target vehicle-mounted image segment, the vehicle controller can write a unique protection identifier to the corresponding image segment. This protection identifier can be associated with the file information of the image segment in the storage device, enabling the storage device's management system to recognize the locked image segment as protected data. During a cyclic overwrite deletion operation, the management system can automatically skip target vehicle-mounted image segments with this protection identifier, only filtering and deleting ordinary vehicle-mounted image segments without protection identifiers. Furthermore, the locking identifier writing operation can be performed in real time after the target vehicle-mounted image segment is identified, and the marking can be directly set in the image file attributes of the storage device. The entire marking process does not require excessive additional system resources and does not affect the normal acquisition and storage of vehicle-mounted images.

[0052] In specific implementations, for example Figure 3 As shown, users can view all locked in-vehicle video clips individually by clicking the "Locked Video" control. Users can also manage stored in-vehicle video clips by clicking the "Video Management" control, allowing them to manage these clips as needed, such as deleting a clip, locking a clip, or determining the severity level of a collision event corresponding to a clip.

[0053] As can be seen, in this embodiment, by first acquiring vehicle images and dividing them into multiple segments for storage, then identifying and locking the image segments related to the vehicle collision, and simultaneously performing cyclic overwriting on the unlocked image segments, it is possible to reasonably utilize the storage space of the storage device to achieve continuous cyclic recording of vehicle images while ensuring that key image segments related to the vehicle collision are effectively preserved and not accidentally deleted. This balances the integrity of the key collision images and the continuity of cyclic recording of vehicle images.

[0054] In one possible embodiment, the method further includes: determining the number of loop recordings of the storage device; if the number of loop recordings is greater than a preset number, determining a reference vehicle video segment from the target vehicle video segment that has been locked; editing the reference vehicle video segment to obtain a edited video segment; locking the edited video segment and deleting the target vehicle video segment that has been locked.

[0055] When determining the number of loop recordings for the storage device, the vehicle controller of the in-vehicle intelligent interactive all-in-one machine can perform full statistics on the loop overwrite operation of the storage device. A single complete first-in-first-out overwrite of an unlocked in-vehicle video clip by the storage device can be counted as one loop recording. The controller can record the cumulative number of loop recordings in real time and store it in the local storage module. It can also update and verify the number of loop recordings according to a preset statistical period to ensure the accuracy of the count. The preset number of recordings can be flexibly set by the user based on the actual capacity of the storage device and the acquisition frequency of the in-vehicle video, and the parameters can be configured through the operation interface of the in-vehicle intelligent interactive all-in-one machine.

[0056] In practice, the reference vehicle-mounted video clips can be video clips marked as processed, video clips whose corresponding collision event severity is lower than a preset level, or a preset number of video clips with the longest locking duration in the storage device. When determining the reference vehicle-mounted video clips, the vehicle controller can retrieve the associated attribute information of all locked and marked video clips, including image acquisition time, collision event association identifier, manual marking information, locking duration data, etc., and then automatically match and filter them according to preset filtering rules.

[0057] As can be seen, this solution first counts the number of loop recordings on the storage device. When the number of loop recordings exceeds a preset value, a reference vehicle video segment is selected from the locked target vehicle video segments and edited. Then, the edited video segment is re-locked and the original locked target vehicle video segment is deleted. This solution can significantly compress the storage volume of locked images while retaining the core image content related to the collision, effectively freeing up the storage space of the storage device and avoiding a large number of locked images occupying storage resources for a long time. This ensures the continuous and stable operation of the vehicle video loop recording function and improves the space utilization of the storage device.

[0058] In one possible embodiment, after determining the reference vehicle image segment from the target vehicle image segment that has been locked, the method further includes: deleting the lock marks from other target vehicle image segments besides the reference vehicle image segment from the target vehicle image segments that have been locked.

[0059] During the unlocking operation, the vehicle controller can first retrieve the file attribute information of all target vehicle image segments that have been locked in the storage device, identify the file identifiers of the vehicle image segments that have been identified as reference vehicle image segments and mark them separately, and then search for the remaining locked target vehicle image segments one by one. The locking mark can be deleted by clearing the corresponding protection identifier in the file attributes of such image segments, or by sending an unlocking command to the image management module of the storage device, so that the management module can modify the locking status of the specified image segment, changing it from a protected locked status to a normal overwriteable status.

[0060] As can be seen, this solution, by identifying the reference vehicle video segment from the locked target vehicle video segment and deleting the locking marks of the remaining locked target vehicle video segments, allows these unlocked video segments to be included in the storage device's loop overlay system. This further frees up long-term occupied storage space, avoids storage resources being squeezed out by too many locked images, thereby improving the space utilization efficiency of the storage device and ensuring the continuous and smooth operation of the vehicle video loop recording function.

[0061] In one possible embodiment, the step of editing the reference vehicle-mounted image segment to obtain an edited image segment includes: determining the collision time from the reference vehicle-mounted image segment; determining the collision time period centered on the collision time; extracting a first collision image segment from the reference vehicle-mounted image segment during the collision time period; analyzing the first collision image segment to obtain an analysis result; and editing the first collision image segment according to the analysis result to obtain an edited image segment.

[0062] Determining the collision time can be achieved by extracting the instantaneous time information from vehicle sensor data associated with a reference vehicle video clip. Alternatively, it can be determined by identifying image features within the reference vehicle video clip. Specifically, the vehicle controller can retrieve gravity sensor monitoring data corresponding to the reference vehicle video clip and directly match the sensor-identified collision instantaneous time information as the collision time for that video clip. Another approach is to analyze the reference vehicle video clip frame by frame, identifying image information such as the moment of contact between the target object and abrupt scene changes to determine the specific frame corresponding to the collision time. Furthermore, the collision time can be cross-checked between sensor data and video image features. This involves identifying the video image based on image features and determining the collision time based on gravity sensor monitoring data, then verifying whether the collision time matches the video image. If they match, the collision time is considered correct.

[0063] Determining the collision time period involves using the precisely located moment of collision as the time center point, extending a preset duration in both the forward and backward directions to define a time interval encompassing the complete scene before, during, and after the collision. The duration of the forward and backward extensions can be flexibly set according to the actual usage scenario of the vehicle, and different time period extension parameters can be configured for different vehicle types. The time period range can also be adjusted based on the collision severity corresponding to reference in-vehicle video footage.

[0064] As can be seen, this solution first accurately determines the collision time from the reference vehicle video footage and then uses this as the center to define the collision time period. After extracting the first collision video footage of this time period, it analyzes it and performs secondary editing based on the analysis results. This allows for the precise editing of collision images by first locking the core image range related to the collision and then removing redundant content. This maximizes the compression of image file size while fully preserving the key image content of the entire collision process, effectively reducing the space occupied by locked images on storage devices and improving the utilization efficiency of storage resources.

[0065] In one possible embodiment, the step of analyzing the first collision image segment to obtain analysis results includes: acquiring the target object and the motion information of the target object included in each frame of the first collision image segment; determining the motion state changes of the target object based on the target object included in each frame of the image and the motion information of the target object; determining multiple target image frames based on the motion state changes, the multiple target image frames indicating the complete collision process; and obtaining analysis results based on the multiple target image frames.

[0066] The extracted target objects can include various collision-related objects such as vehicles, pedestrians, and obstacles. Determining the motion state changes of the target objects involves temporally comparing the motion information of the same target object in consecutive frames, analyzing the continuous changes in displacement, velocity, direction of motion, and attitude during the collision period. It also includes identifying the full-stage characteristic changes of the target object from normal motion to contact, collision, and post-collision state changes. Specifically, the motion information of the target object extracted from each frame, such as position coordinates, trajectory, angular velocity, and attitude angle, can be sequentially organized into a motion data sequence. By calculating the difference in motion data between adjacent frames, it can be determined whether the target object's motion state is uniform, accelerating, decelerating, or undergoing a sudden change. Simultaneously, it can identify whether the target object comes into contact with other objects and the morphological changes after contact, among other collision-related state changes. Furthermore, it can analyze the relative motion state changes between multiple target objects to clarify the interaction relationships between different target objects during the collision process, ensuring complete capture of all key state changes of the target objects before and after the collision.

[0067] Determining multiple target image frames based on changes in motion state involves selecting image frames from consecutive frames of the first collision image segment that fully present key changes in the target object's motion state. Specifically, this can begin by locating image frames corresponding to critical time points where the target object's motion state undergoes abrupt changes, such as the initial frame where the target object begins to decelerate before the collision, the contact frame where the target objects first make contact, the core frame where the morphological changes are most obvious during the collision, and the ending frame where the target object's motion tends to stabilize after the collision. Alternatively, representative image frames can be selected from the key state stages of the entire collision process at preset frame intervals. This ensures that the selected multiple target image frames continuously and completely cover the cause, process, and result of the collision in the temporal dimension, without omitting key collision scenes or selecting too many redundant frames without significant state changes.

[0068] When determining the target image frame, the selected frame can be verified a second time to confirm that it can clearly reflect the motion state change characteristics of the target object. The number of target image frames can also be flexibly adjusted according to the frame rate of the vehicle image, so as to control the rationality of the number of frames while ensuring the complete presentation of the collision process.

[0069] As can be seen, this solution acquires the target object and motion information in the first collision image segment frame by frame, analyzes the changes in its motion state, and determines multiple target image frames that can indicate the complete collision process. Based on this, the analysis results can be obtained, which can accurately identify the core image content of the collision process, provide accurate content basis for subsequent editing, achieve fine screening of collision images, retain the key scenes of the entire collision process while removing redundant image frames, maximize the compression of image file size, and reduce the space occupation of locked images on storage devices.

[0070] In one possible embodiment, determining whether there is a target vehicle image segment related to a vehicle collision among the plurality of vehicle image segments includes: acquiring gravity monitoring information from the gravity sensor of the target vehicle; determining target gravity monitoring information from the gravity monitoring information, wherein the target gravity monitoring information is the gravity detection information at the moment of the collision of the target vehicle; determining the target time for acquiring the target gravity monitoring information; and determining the vehicle image segment among the plurality of vehicle image segments that includes the target time as the target vehicle image segment.

[0071] When acquiring gravity monitoring information from the gravity sensor, the vehicle controller can establish a data communication connection with the gravity sensor via the vehicle's Controller Area Network (CAN) bus. The gravity sensor can collect gravity change data in real time during vehicle operation and transmit the monitoring information to the vehicle controller at a preset frequency. The transmitted gravity monitoring information may include timestamps, raw data collected by the sensor, etc. When determining the target gravity monitoring information from the gravity monitoring information, the vehicle controller can perform real-time analysis and filtering of the cached gravity monitoring information, identifying the gravity detection information that matches the instant of vehicle collision from the continuous monitoring data and determining it as the target gravity monitoring information.

[0072] As can be seen, this solution obtains the gravity monitoring information from the vehicle's gravity sensor and determines the target gravity monitoring information at the moment of collision. Then, it locks the corresponding target time and matches the vehicle image segment containing that time as the target vehicle image segment. It can achieve rapid and accurate positioning of collision-related image segments by relying on the precise monitoring data of the sensor, making the judgment of key collision image segments more objective and accurate, and effectively avoiding missed or wrong judgments of collision-related vehicle image segments.

[0073] In one possible embodiment, determining the target gravity monitoring information from the gravity monitoring information includes: acquiring the three-axis acceleration information of the target vehicle based on the gravity monitoring information, wherein the three axes indicate two horizontal directions and one vertical direction; generating time-acceleration waveform data for each axis based on the three-axis acceleration information; acquiring the vehicle type of the target vehicle; acquiring the driving scenario of the target vehicle in each time period; and determining the target gravity monitoring information based on the vehicle type, the driving scenario, and the time-acceleration waveform data for each axis.

[0074] The three-axis acceleration includes the lateral acceleration in the horizontal direction, the longitudinal acceleration in the horizontal direction, and the vertical acceleration in the vertical direction during vehicle movement. These three sets of acceleration data can fully reflect the force changes of the vehicle in three-dimensional space. The on-board controller can extract the raw acceleration data from the gravity monitoring information transmitted by the gravity sensor and convert it into standardized three-axis acceleration information through a preset parsing algorithm.

[0075] The generation of time-acceleration waveform data involves arranging the acceleration information for each axis according to the acquisition timestamps, with time as the horizontal axis and acceleration value as the vertical axis, transforming continuous acceleration data into a visualized waveform curve. During the generation process, interpolation can be performed on the time-series acceleration data to ensure the continuity and smoothness of the waveform. Acceleration data can be sampled and integrated at preset time intervals to adapt to different data analysis precision requirements.

[0076] Vehicle types include various categories such as small passenger cars, medium-sized buses, heavy commercial vehicles, and trucks. The onboard controller can directly obtain preset vehicle type data by reading the vehicle's basic information, or the user can manually enter the vehicle type information through the onboard intelligent interactive interface. Determining the driving scenario involves comprehensively judging the vehicle's driving environment through the vehicle's positioning and navigation module, environmental perception sensors, and driving status data. Determinable scenarios include urban paved roads, rural unpaved roads, highways, bumpy mountain roads, and congested urban sections. The onboard controller can obtain the vehicle's location information through positioning information and combine it with electronic map data to initially determine the driving scenario. It can also use driving data collected by the vehicle's speed sensor and bump detection sensor, as well as road environment images collected by the onboard camera, for secondary verification and precise determination of the driving scenario. Furthermore, the determined driving scenario can be dynamically updated according to a preset time period to ensure that the driving scenario data in each time period matches the actual driving status of the vehicle.

[0077] As can be seen, this solution extracts the vehicle's three-axis acceleration information from gravity monitoring data and generates time-acceleration waveform data for each axis. Then, it combines the acquired vehicle type and driving scenarios at different times to comprehensively analyze and determine the target gravity monitoring information. This approach can accurately determine the gravity monitoring information at the moment of collision by combining the vehicle's own attributes and the actual driving environment. It effectively avoids acceleration data interference caused by different vehicle models and driving scenarios, improves the accuracy and adaptability of the target gravity monitoring information determination, and thus provides reliable data support for the accurate positioning of collision-related image fragments.

[0078] In one possible embodiment, determining the target gravity monitoring information based on the vehicle type, the driving scenario, and the time-acceleration waveform data for each axle includes: determining waveform characteristics for each time period based on the time-acceleration waveform data for each axle, wherein the waveform characteristics include at least one of peak acceleration, waveform rise slope, or waveform duration; determining waveform characteristic thresholds for each axle in the three-axis acceleration information based on the vehicle type and the driving scenario; and determining the target gravity monitoring information based on the waveform characteristics for each time period and the waveform characteristic thresholds for each axle.

[0079] The determination of target gravity monitoring information based on waveform characteristics within each time period and waveform characteristic thresholds for each axis includes comparing the waveform characteristic values ​​of each axis within each time period with the corresponding waveform characteristic thresholds for that axis, filtering out time periods corresponding to waveform characteristics whose values ​​exceed the thresholds, and initially marking the gravity monitoring information within those time periods as collision-related monitoring information. The comparison results of the three axes are comprehensively verified, and the threshold exceedance of multi-axis waveform characteristics is used to determine whether an actual collision has occurred. Finally, the gravity monitoring information corresponding to the moment of collision is determined to be the target gravity monitoring information.

[0080] In practice, the vehicle controller first performs threshold comparisons on the peak acceleration, waveform rise slope, and waveform duration for the lateral, longitudinal, and vertical axes. If any waveform feature of one or more axes exceeds the corresponding threshold, the monitoring information for that time period is listed as candidate information. Then, a second round of screening is conducted using preset comprehensive judgment rules, such as multiple axis features simultaneously exceeding thresholds or core features (such as peak acceleration) exceeding thresholds by a preset proportion. This excludes acceleration abrupt changes caused by non-collision factors such as road bumps and sudden braking. Simultaneously, the screened collision-time monitoring information can be precisely time-localized, matched with the corresponding gravity sensor data, and completely extracted and marked as target gravity monitoring information. The vehicle controller can segment and analyze the generated time-acceleration waveform data. It can divide the analysis time period according to preset time intervals or dynamically segment it based on the waveform's changing trend. During the analysis process, a dedicated waveform feature extraction algorithm can accurately obtain characteristic values ​​such as peak acceleration, waveform rise slope, and waveform duration for each time period.

[0081] When determining the waveform characteristic thresholds for each axle based on vehicle type and driving scenario, the onboard controller can retrieve a locally stored threshold configuration library. This library pre-stores three-axis waveform characteristic threshold data for different vehicle types under various driving scenarios. For example, the peak vertical acceleration threshold for a heavy commercial vehicle on bumpy mountain roads will be higher than the corresponding threshold for a small passenger car on paved urban roads. The onboard controller can accurately match the corresponding thresholds for each waveform characteristic of each axle from the configuration library based on the acquired vehicle type and real-time driving scenario. Users can also personalize the thresholds through the onboard intelligent interactive interface according to their actual needs. The adjusted thresholds can be synchronized to the data processing module in real time to adapt to different usage scenarios and judgment requirements.

[0082] As can be seen, this solution extracts waveform features such as peak acceleration from the time-acceleration waveform data of each axis, matches the waveform feature thresholds corresponding to each axis with the vehicle type and driving scenario, and then compares the waveform features with the thresholds to determine the target gravity monitoring information. It can accurately determine the gravity monitoring information at the moment of collision through quantitative feature threshold comparison, effectively eliminate the acceleration data interference caused by non-collision factors under different vehicle models and driving scenarios, and greatly improve the accuracy and reliability of the target gravity monitoring information determination.

[0083] In one possible embodiment, determining whether there is a target vehicle image segment related to a vehicle collision among the plurality of vehicle image segments includes: obtaining an emergency locking command; determining the locking time of the emergency locking command; and obtaining a target vehicle image segment based on the locking time, wherein the target vehicle image segment includes the vehicle image at the locking time.

[0084] Among them, for example Figure 4 As shown, users can initiate emergency locking of the current real-time video segment by clicking the "Emergency Lock" control or by using voice commands. When determining the video segment to be urgently locked, the system checks whether the in-vehicle video after receiving the emergency lock command includes footage corresponding to the collision. If it does, the in-vehicle video segment after receiving the emergency lock command is identified as the target segment. If not, the system analyzes and determines the video footage within a preset time period prior to receiving the emergency lock command, identifying the footage within that preset time period that includes footage corresponding to the collision, and then selecting the footage containing the collision-related information as the target segment.

[0085] In practice, after emergency locking, the system can again capture the user's click on the "unlock" control to determine if the image before receiving the unlock command is part of the target vehicle video segment. That is, the target vehicle video segment includes images of the vehicle from the moment the emergency locking command is received until a preset time period, up to the moment the unlock command is received. If no unlock command is received within a relatively long preset time period, the lock is automatically released.

[0086] As can be seen, this solution obtains the emergency locking command and determines the locking time, and then matches the vehicle image segment containing the locking time as the target vehicle image segment. This enables the accurate locking of collision-related image segments triggered manually, making up for the limitations of automatic sensor judgment. It can flexibly capture various types of driving images that need to be retained, improve the flexibility and comprehensiveness of the determination of target vehicle image segments, and ensure that important driving images are not missed.

[0087] This application also provides a loop recording device for an in-vehicle intelligent interactive all-in-one machine, comprising: an acquisition unit for acquiring in-vehicle images of a target vehicle; a division unit for dividing the in-vehicle images into multiple in-vehicle image segments and storing the multiple in-vehicle image segments in a storage device; a determination unit for determining whether there is a target in-vehicle image segment related to a vehicle collision among the multiple in-vehicle image segments; and a locking marking unit for locking the target in-vehicle image segment if it exists, wherein in-vehicle image segments that are not locked are cyclically overwritten in the storage device.

[0088] Please see Figure 5 , Figure 5 This is a structural schematic diagram of the in-vehicle intelligent interactive all-in-one machine provided in this application. (See attached diagram.) Figure 5 As shown, the in-vehicle intelligent interactive all-in-one machine may include: an in-vehicle controller 510, a communication interface 520, a storage device 530, and a communication bus 540. The in-vehicle controller 510, communication interface 520, and storage device 530 communicate with each other via the communication bus 540. The in-vehicle controller 510 can call logical instructions in the storage device 530 to perform the following operations: acquire in-vehicle images of the target vehicle; divide the in-vehicle images into multiple in-vehicle image segments and store the multiple in-vehicle image segments in the storage device; determine whether there is a target in-vehicle image segment related to a vehicle collision among the multiple in-vehicle image segments; if the target in-vehicle image segment exists, lock the target in-vehicle image segment, wherein in-vehicle image segments that are not locked will be cyclically overwritten in the storage device.

[0089] Furthermore, the logical instructions in the aforementioned storage device 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0090] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a loop recording method for an in-vehicle intelligent interactive all-in-one machine provided by the above methods. The method includes: acquiring an in-vehicle image of a target vehicle; dividing the in-vehicle image into multiple in-vehicle image segments and storing the multiple in-vehicle image segments in a storage device; determining whether there is a target in-vehicle image segment related to a vehicle collision among the multiple in-vehicle image segments; if the target in-vehicle image segment exists, locking the target in-vehicle image segment, wherein in-vehicle image segments that are not locked are cyclically overwritten in the storage device.

[0091] In another aspect, this application also provides a computer program product, including a computer program that, when executed by a processor, implements a loop recording method for any of the above-described in-vehicle intelligent interactive all-in-one machines. The method includes: acquiring in-vehicle images of a target vehicle; dividing the in-vehicle images into multiple in-vehicle image segments and storing the multiple in-vehicle image segments in a storage device; determining whether a target in-vehicle image segment related to a vehicle collision exists among the multiple in-vehicle image segments; if the target in-vehicle image segment exists, locking the target in-vehicle image segment, wherein in-vehicle image segments that are not locked are cyclically overwritten in the storage device.

[0092] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0093] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

Claims

1. A method for loop recording of an in-vehicle intelligent interactive all-in-one machine, characterized in that, include: Acquire in-vehicle images of the target vehicle; The vehicle image is divided into multiple vehicle image segments, and the multiple vehicle image segments are stored in a storage device; Determine whether any of the plurality of vehicle image segments are related to a vehicle collision; If the target vehicle image segment exists, the target vehicle image segment is locked. Vehicle image segments that are not locked are cyclically overwritten in the storage device.

2. The method according to claim 1, characterized in that, The method further includes: Determine the number of loop recordings for the storage device; If the number of recorded cycles exceeds a preset number, a reference vehicle video segment is determined from the target vehicle video segments that are marked as locked. The reference vehicle-mounted image segment is edited to obtain an edited image segment; The edited video clip is locked and then deleted.

3. The method according to claim 2, characterized in that, After determining the reference vehicle image segment from the target vehicle image segment marked with a lock, the method further includes: Remove the lock marks from the target vehicle image segments that are marked as locked, except for the reference vehicle image segment.

4. The method according to claim 2, characterized in that, The process of editing the reference vehicle-mounted image segment to obtain an edited image segment includes: The collision time is determined from the reference vehicle image footage; The collision time period is determined with the collision time as the center. Extract a first collision image segment from the reference vehicle image segment during the collision period; The first collision image segment was analyzed to obtain the analysis results; Based on the analysis results, the first collision image segment is edited to obtain an edited image segment.

5. The method according to claim 4, characterized in that, The analysis of the first collision image segment to obtain the analysis results includes: Obtain the target object and its motion information in each frame of the first collision image segment; The motion state changes of the target object are determined based on the target object included in each frame of image and the motion information of the target object. Multiple target image frames are determined based on the changes in the motion state, and the multiple target image frames indicate the complete collision process; Analysis results are obtained based on the multiple target image frames.

6. The method according to claim 1, characterized in that, Determining whether there is a target vehicle image segment related to a vehicle collision among the plurality of vehicle image segments includes: Obtain gravity monitoring information from the gravity sensor of the target vehicle; The target gravity monitoring information is determined from the gravity monitoring information, wherein the target gravity monitoring information is the gravity detection information at the moment of collision of the target vehicle; Determine the target time for acquiring the target gravity monitoring information; The vehicle image segment that includes the target time among the plurality of vehicle image segments is identified as the target vehicle image segment.

7. The method according to claim 6, characterized in that, Determining the target gravity monitoring information from the gravity monitoring information includes: The three-axis acceleration information of the target vehicle is obtained based on the gravity monitoring information, where the three axes indicate two horizontal directions and one vertical direction; Generate time-acceleration waveform data for each axis based on the triaxial acceleration information; Obtain the vehicle type of the target vehicle; Obtain the driving scene of the target vehicle within each time period; Target gravity monitoring information is determined based on the vehicle type, the driving scenario, and the time-acceleration waveform data for each axle.

8. The method according to claim 7, characterized in that, The step of determining target gravity monitoring information based on the vehicle type, the driving scenario, and the time-acceleration waveform data for each axle includes: The waveform characteristics within each time period are determined based on the time-acceleration waveform data of each axis, and the waveform characteristics include at least one of peak acceleration, waveform rise slope, or waveform duration. The waveform feature threshold for each axis in the three-axis acceleration information is determined based on the vehicle type and the driving scenario. The target gravity monitoring information is determined based on the waveform characteristics within each time period and the waveform characteristic threshold for each axis.

9. The method according to claim 1, characterized in that, Determining whether there is a target vehicle image segment related to a vehicle collision among the plurality of vehicle image segments includes: Obtain emergency locking command; Determine the locking time of the emergency locking command; The target vehicle image segment is obtained based on the locking time, and the vehicle image segment includes the vehicle image at the locking time.

10. A vehicle-mounted intelligent interactive all-in-one machine, characterized in that, Includes an on-board controller, the on-board controller being used for: Acquire in-vehicle images of the target vehicle; The vehicle image is divided into multiple vehicle image segments, and the multiple vehicle image segments are stored in a storage device; Determine whether any of the plurality of vehicle image segments are related to a vehicle collision; If the target vehicle image segment exists, the target vehicle image segment is locked. Vehicle image segments that are not locked are cyclically overwritten in the storage device.

Images