A driving scene recognition method, an electronic device, and a storage medium

Abstract

The application provides a driving scene recognition method, an electronic device and a storage medium, and relates to the technical field of vehicles. The driving scene recognition method provided by the application comprises the following steps: acquiring multi-source data related to a driving trip in the process of driving a vehicle, wherein the multi-source data comprises space-time data and environment perception data of the vehicle; determining a driving scene according to the space-time data and the environment perception data; and displaying a trip review report corresponding to the driving scene on a preset interface according to the multi-source data and the driving scene. The application can accurately recognize the driving scene without relying on a single data source, and can generate a trip review report by combining the multi-source data and the driving scene after the trip ends, so that the driving trip can be sorted and presented in a scene-based manner, thereby improving the emotional value of the user's driving experience.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and more specifically, to a driving scene recognition method, electronic device, and storage medium. Background Technology

[0002] During operation, vehicles can continuously collect multi-source data, including location information, time information, vehicle attitude information, and images from onboard cameras. Based on this data, vehicle systems or onboard applications generally have driving trip recording and statistical functions to display basic information such as single or periodic driving mileage, driving time, average speed, and route trajectory to users.

[0003] Currently, some manufacturers have achieved a certain degree of automation and visualization in trip statistics, such as summarizing driving behavior through map trajectory playback and trip data reports. However, the focus of these existing technologies is mainly on the objective quantitative statistics of driving behavior. Their technical implementation usually revolves around basic parameters such as mileage, time, and speed, lacking further subdivision and identification of the specific environment and scenarios during the driving process. Summary of the Invention

[0004] The problem addressed in this application is how to achieve scene recognition and review.

[0005] To address the aforementioned issues, this application provides a driving scene recognition method, an electronic device, and a storage medium.

[0006] Firstly, this application provides a driving scene recognition method, including: During vehicle operation, multi-source data related to the driving journey is acquired, including the vehicle's spatiotemporal data and environmental perception data. The driving scenario is determined based on the spatiotemporal data and the environmental perception data. Based on the multi-source data and the driving scenario, a trip review report corresponding to the driving scenario is displayed on a preset interface.

[0007] Optionally, the driving scenario includes a celestial scenario, and determining the driving scenario based on the spatiotemporal data and the environmental perception data includes: Based on the spatiotemporal data, determine the theoretical spatial position information of the celestial target corresponding to the celestial scene relative to the vehicle; Based on the environmental perception data, candidate visual targets corresponding to the celestial targets are detected; The spatial location information of the candidate visual target is matched and verified based on the theoretical spatial location information. When the matching verification is successful, it is determined that there is a star scene corresponding to the star target.

[0008] Optionally, the theoretical spatial location information includes the elevation angle and azimuth angle of the celestial target, and the matching and verification of the spatial location information of the candidate visual target based on the theoretical spatial location information includes: The spatial location information of the candidate visual target is matched and verified with the theoretical spatial location information of the celestial target. When the difference between the elevation angle and azimuth angle of the candidate visual target and the elevation angle and azimuth angle of the celestial target is less than a preset angle threshold, the matching verification is deemed successful.

[0009] Optionally, determining the driving scenario based on the spatiotemporal data and the environmental perception data further includes: The stability of the scale changes of the candidate visual targets over a continuous time period is verified; Once the matching verification and stability verification are passed, it is determined that there is a star scene corresponding to the star target.

[0010] Optionally, the stability verification of the scale changes of the candidate visual target over a continuous time period includes: The candidate visual target is tracked in multiple consecutive image frames to obtain the scale parameters of the candidate visual target in each image frame; The scale change rate of the candidate visual target over a continuous time period is calculated based on the scale parameter. In response to determining that the candidate visual target satisfies the scale stability condition based on the scale change rate, the stability verification is deemed successful.

[0011] Optionally, determining the driving scenario based on the spatiotemporal data and the environmental perception data includes: When the driving scenario includes a weather scenario, the weather triggering conditions corresponding to the weather scenario are determined based on the spatiotemporal data; when the vehicle's position meets the weather triggering conditions, candidate visual features corresponding to the weather scenario are detected based on the environmental perception data; when the candidate visual features meet the preset weather scenario feature requirements, it is determined that the weather scenario exists. And / or, When the driving scenario includes a landscape scenario, the spatial relationship between the vehicle's position and a preset landscape area is determined based on the spatiotemporal data; if the spatial relationship is satisfied, candidate visual features corresponding to the landscape scenario are detected based on the environmental perception data; when the candidate visual features meet the preset landscape scenario feature requirements, it is determined that the landscape scenario exists.

[0012] Optionally, before displaying the trip review report corresponding to the driving scenario on a preset interface based on the multi-source data and the driving scenario, the method further includes: The scenario-specific data corresponding to the driving scenario is linked and integrated with the basic trip data to generate scenario-specific trip data; A trip review report corresponding to the driving scenario is generated based on the scenario-based trip data.

[0013] Optionally, the trip review report corresponding to the driving scenario includes at least one of the following: scenario segment data, timeline data, scenario location information, and scenario summary information.

[0014] Secondly, this application provides an electronic device, including a memory and a processor; The memory is used to store computer programs; The processor is configured to implement the driving scene recognition method as described in the first aspect when executing the computer program.

[0015] Thirdly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the driving scene recognition method as described in the first aspect.

[0016] The beneficial effects of the driving scene recognition method in this application are: by acquiring multi-source data related to the driving trip during the vehicle's operation, and making a comprehensive judgment on the driving scene based on spatiotemporal data and environmental perception data, the method can accurately identify the driving scene without relying on a single data source. After the trip, the method combines multi-source data and the driving scene to generate a trip review report, which enables the driving trip to be organized and presented in a scenario-based manner, thereby enhancing the emotional value of the user's driving experience. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the driving scene recognition method according to an embodiment of this application; Figure 2 This is a schematic diagram of the process for determining the driving scenario in an embodiment of this application. Figure 1 ; Figure 3 This is a schematic diagram of the process for determining the driving scenario in an embodiment of this application. Figure 2 ; Figure 4 This is a schematic diagram of the stability verification process in an embodiment of this application; Figure 5 This is a schematic diagram of the process for determining the driving scenario in an embodiment of this application. Figure 3 ; Figure 6 This is a schematic diagram of the process for determining the driving scenario in an embodiment of this application. Figure 4 ; Figure 7This is a flowchart illustrating the itinerary review report in an embodiment of this application; Figure 8 This is a system architecture diagram of the driving scene recognition device according to an embodiment of this application; Figure 9 This is a system architecture diagram of an electronic device according to an embodiment of this application; Figure 10 This is an example diagram of a trip review report according to an embodiment of this application. Detailed Implementation

[0018] To make the above-mentioned objects, features, and advantages of this application more apparent and understandable, specific embodiments of this application are described in detail below with reference to the accompanying drawings. Although some embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the accompanying drawings and embodiments of this application are for illustrative purposes only and are not intended to limit the scope of protection of this application.

[0019] It should be understood that the steps described in the method embodiments of this application may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this application is not limited in this respect.

[0020] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the description below. It should be noted that the concepts of "first," "second," etc., mentioned in this application are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0021] It should be noted that the terms "one" and "more" used in this application are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices in the embodiments of this application are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0023] like Figure 1As shown in the embodiment of this application, a driving scene recognition method includes: S100: During vehicle operation, acquire multi-source data related to the driving journey, wherein the multi-source data includes the vehicle's spatiotemporal data and environmental perception data.

[0024] Specifically, during vehicle operation, multi-source data related to the driving journey is acquired through multi-source data acquisition modules such as vehicle positioning module (e.g., vehicle GPS), time synchronization module (e.g., time synchronization module), and environmental perception device (e.g., forward-facing camera). For example, the vehicle's location information (e.g., latitude and longitude) and speed data are acquired through vehicle GPS, world standard time is acquired through time synchronization module, real-time forward-facing images of the vehicle are acquired through forward-facing camera, and pitch angle, azimuth angle, and roll angle of the vehicle are acquired through inertial measurement unit.

[0025] Spatiotemporal data refers to data used to characterize the spatial location of a vehicle during a driving journey and its relationship with time. It is used to reflect the vehicle's geographical location, driving status, and time-related spatial distribution characteristics at different points in time. Spatiotemporal data includes time-related data (such as the start time and end time of the journey), spatial location-related data (such as satellite positioning system positioning data, latitude and longitude coordinate data, altitude information, etc.), driving trajectory and route data, location association information (such as the relative relationship between the vehicle's position and the road, terrain, or regional boundary), and derived spatiotemporal parameters (such as driving speed, azimuth angle, etc. calculated based on the above).

[0026] Among them, environmental perception data refers to the data acquired by the vehicle through environmental perception devices to characterize the characteristics of the surrounding environment during the vehicle's operation. It can reflect the perceptible objects, scenes, or environmental state information in the external environment of the vehicle. Environmental perception data includes visual perception data (such as image data collected by vehicle cameras), environmental target perception data (such as the position parameters of the target in the image, the size, area, or scale parameters of the target, etc.), environmental state perception data (such as light intensity information, visibility information, etc.), and multi-sensor perception data (such as radar perception data, lidar point cloud data, and ultrasonic perception data).

[0027] S200: Determine the driving scenario based on the spatiotemporal data and the environmental perception data.

[0028] Specifically, the driving scenarios during vehicle operation are analyzed and judged based on spatiotemporal data and environmental perception data to determine whether a preset driving scenario exists in the current driving journey.

[0029] Among them, a driving scenario refers to a driving situation with identifiable characteristics that is jointly constituted by the spatiotemporal state of the vehicle and the characteristics of the surrounding environment during the vehicle's operation. A driving scenario is used to characterize the driving state or driving experience of a vehicle at a specific time, at a specific location, and under specific environmental conditions. Exemplary driving scenarios include celestial scenarios, meteorological scenarios, and landscape scenarios. It may also include other driving scenarios with identifiable characteristics determined based on the vehicle's driving time, driving location, environmental state, or environmental content. The specific scenario types are not limited to the scenarios listed in this embodiment.

[0030] S300: Based on the multi-source data and the driving scenario, display a trip review report corresponding to the driving scenario on a preset interface.

[0031] Specifically, after the trip, based on multi-source data and the determined driving scenario, a trip review report corresponding to the driving scenario is generated and displayed on a preset interface. Taking the driving scenario as a sunset car (belonging to the star scene) as an example, the corresponding trip review report can be generated by integrating the start and end times, latitude and longitude, and corresponding image frames of the sunset car. For example, scene segment data (video segments generated from continuous environmental image data corresponding to the sunset car scenario, reflecting the driving scene with a stable relative position of the sun and obvious color changes during the vehicle's driving process), time axis data (marking the time interval of the sunset car scenario in the trip time axis), scene location information (marking the driving route location corresponding to the sunset car scenario on the map), and scene summary information (scene description information generated based on the duration of the sunset car scenario and the type of driving section, such as "12-minute national highway sunset car trip").

[0032] In this embodiment, by acquiring multi-source data related to the driving trip during vehicle operation, and making a comprehensive judgment on the driving scenario based on spatiotemporal data and environmental perception data, the driving scenario can be accurately identified without relying on a single data source. After the trip, a trip review report is generated by combining multi-source data and driving scenario, so that the driving trip can be organized and presented in a scenario-based manner, thereby enhancing the emotional value of the user's driving experience.

[0033] Optionally, the driving scenario includes a celestial scenario, and determining the driving scenario based on the spatiotemporal data and the environmental perception data includes: S110A: Determine the theoretical spatial position information of the celestial target corresponding to the celestial scene relative to the vehicle based on the spatiotemporal data.

[0034] Specifically, in combination Figure 2As shown, the driving scenario includes a star scene. Based on the spatiotemporal data acquired during vehicle operation, the theoretical spatial position information of the star target corresponding to the star scene relative to the vehicle at the current moment is calculated, such as the theoretical directional position of the star target in the sky.

[0035] Among them, a star scene refers to a driving scene related to a celestial target that is determined during vehicle operation based on the vehicle's time and geographical location information, combined with the theoretical spatial position information of celestial bodies at the corresponding time, and verified by environmental perception data to verify the visual characteristics of celestial bodies. In other words, a driving scene can be determined as a star scene, which usually meets at least one or more of the following technical characteristics: (1) There is a clear celestial target: such as the sun, moon, or a bright star that can be visually perceived (such as Venus, Jupiter, etc.); (2) The celestial target has calculable theoretical spatial position information; (3) There is a corresponding celestial target. The environmental visual characteristics; (4) The theoretical spatial location information and the visual detection results satisfy the matching relationship, that is, the spatial location of the candidate visual target in the image is consistent with or close to the theoretical elevation angle and azimuth angle of the celestial target; Exemplary star scenes include: sunrise / sunset related driving scenes, moon related driving scenes, bright star or starry sky related driving scenes; It should be noted that if only bright spots are detected in the sky, but cannot be matched with the theoretical celestial position, or if artificial light sources such as car lights and street lights appear in the environmental image and affect the brightness of the celestial body, or if the celestial body is severely obscured and visual verification cannot be completed, it should generally not be identified as a star scene.

[0036] Among them, celestial targets refer to celestial bodies or celestial phenomena that are identified as objects in celestial scenes. Celestial targets have theoretical spatial position information that can be calculated based on the time and geographical location of the vehicle, and can form detectable visual features in environmental perception data. That is, an object can be identified as a celestial target, usually having one or more of the following technical features: (1) having celestial attributes: celestial targets can be natural celestial bodies or celestial phenomena, including but not limited to: the sun, the moon, bright stars that can be perceived visually, and celestial phenomena formed during the rising or setting of celestial bodies; (2) having theoretical spatial position information: celestial targets have an altitude angle and azimuth angle relative to the vehicle position at any given time; (3) having detectability in environmental perception. It should be noted that artificial light sources such as car lights, street lights, and advertising signs, or non-celestial light sources such as aircraft lights and reflected light, or random bright spots that cannot be correlated with theoretical celestial positions are usually not celestial targets.

[0037] S120A: Detect candidate visual targets corresponding to the celestial targets based on the environmental perception data.

[0038] Specifically, based on environmental perception data, the image of the environment in front of the vehicle is analyzed to detect candidate visual targets corresponding to celestial targets.

[0039] S130A: The spatial location information of the candidate visual target is matched and verified according to the theoretical spatial location information. When the matching verification is successful, it is determined that there is a star scene corresponding to the star target.

[0040] Specifically, the spatial location information of the candidate visual target in the image is matched and verified with the theoretical spatial location information of the star target. When the two meet the preset matching conditions, it is determined that there is a star scene corresponding to the star target in the current driving journey.

[0041] In this optional embodiment, the theoretical spatial location information of the celestial target is calculated based on spatiotemporal data, and candidate visual targets are detected and matched for verification by combining environmental perception data. This can utilize the objective physical laws of celestial positions to identify celestial scenes, effectively avoiding the false detection problem caused by relying solely on image features, thereby improving the accuracy and reliability of celestial scene recognition.

[0042] Optionally, the theoretical spatial location information includes the elevation angle and azimuth angle of the celestial target, and the matching and verification of the spatial location information of the candidate visual target based on the theoretical spatial location information includes: The spatial location information of the candidate visual target is matched and verified with the theoretical spatial location information of the celestial target. When the difference between the elevation angle and azimuth angle of the candidate visual target and the elevation angle and azimuth angle of the celestial target is less than a preset angle threshold, the matching verification is deemed successful.

[0043] Specifically, the theoretical spatial location information includes the elevation angle and azimuth angle of the celestial target. The location prediction module can calculate the theoretical elevation angle (e.g., solar elevation angle hs) and theoretical azimuth angle (e.g., solar azimuth angle As) of the celestial target at the current moment based on the vehicle's location information and time information. At the same time, based on the positional relationship of the candidate visual target in the image, the module calculates the elevation angle and azimuth angle corresponding to the candidate visual target. When the difference between the elevation angle of the candidate visual target and the theoretical elevation angle of the celestial target, and the difference between the azimuth angle of the candidate visual target and the theoretical azimuth angle of the celestial target are both less than a preset angle threshold (e.g., 5°), the matching verification is deemed successful.

[0044] In determining the theoretical spatial location information of a celestial target (such as the Sun), the solar declination δ and solar hour angle Ts can also be determined, where δ = 23.45 × sin[360 × (284 + n) / 365], n is the accumulated days in a year, and Ts = (true solar time - 12) × 15°.

[0045] Among these, the solar altitude angle hs also needs to be corrected for atmospheric refraction, i.e., the corrected solar altitude angle hs 修正 =hs+0.0167° / tan(hs+0.5667°).

[0046] In this optional embodiment, the theoretical spatial location information of the celestial target is specifically defined as the elevation angle and azimuth angle, and the candidate visual target is matched and verified based on a preset angle threshold. This can transform the celestial scene recognition process into a quantifiable angle matching process, making the scene recognition standard clearer, thereby improving the consistency and repeatability of the recognition results.

[0047] Optionally, determining the driving scenario based on the spatiotemporal data and the environmental perception data further includes: S140: Verify the stability of the scale changes of the candidate visual target over a continuous time period.

[0048] Specifically, in combination Figure 3 As shown, while performing matching verification, the stability of the scale change of the candidate visual target in a continuous time period is also verified. For example, during the continuous driving of the vehicle, multiple frames of environmental images are continuously acquired, the candidate visual target is continuously tracked, and its scale change characteristics in a continuous time period are analyzed.

[0049] S150: After the matching verification and stability verification are passed, it is determined that there is a star scene corresponding to the star target.

[0050] Specifically, when a candidate visual target passes both the matching verification and the stability verification, it is determined that there is a star scene corresponding to the star target in the current driving journey.

[0051] Among them, celestial targets are the objects to be identified in celestial scenes. A celestial scene is a driving situation formed around a celestial target under specific spatiotemporal conditions. One celestial target can correspond to one or more celestial scenes. That is, the same celestial target can correspond to different celestial scenes under different time conditions, spatial direction conditions, or vehicle driving states. For example, the sun can correspond to sunrise driving scene, sunset driving scene, and sunset flying scene, and the moon can correspond to moon chasing driving scene and bright moon accompanying driving scene, etc. The difference between different celestial scenes is not only in the celestial target itself, but also in the spatiotemporal relationship and relative motion relationship between the celestial target and the vehicle. For example, the sun before and after sunrise corresponds to sunrise driving scene, the moon during moonrise corresponds to moon chasing driving scene, and the vehicle driving towards the celestial target can correspond to sun chasing / moon chasing driving scene. A celestial target that is stable in the field of vision for a long time can correspond to accompanying driving scene.

[0052] For example, taking the sunset car chase scene as an example, the local sunset time can be obtained by querying a pre-stored astronomical database based on the vehicle's location information (e.g., latitude and longitude). If the current time is one hour before sunset, and the vehicle speed is between 10 km / h and 350 km / h, and the acceleration a ≤ 0.3g, then the time verification of the sunset car chase scene is passed. Then, the real-time forward-facing image of the vehicle obtained by the forward-facing camera can be identified. First, the image is preprocessed, and Gaussian filtering (3×3 kernel) and adaptive threshold segmentation (red channel priority) are used to extract high-brightness areas. Then, the circularity (≥0.9) is calculated by Hough circle transform (minimum radius 20px / maximum radius 200px) to filter candidate sun areas. Finally, the image detection stage is completed by camera coordinate transformation, which can find circular targets that "may be the sun" in the image. However, the image alone cannot distinguish between the real sun and the red circle that imitates the shape of the sun. The system first identifies obstructions such as streetlights and billboards. Then, it performs matching and stability verification in sequence. Matching verification effectively eliminates objects that resemble the sun but appear in the wrong sky position (such as a round red decorative light installed on the roadside). At this point, the sun can be identified with a high probability, but it may still be fooled by a round light source of fixed size and coincident position in the distance. Therefore, stability verification is also required. Considering that the scale of objects at close range changes rapidly in the image, while the sun, as a distant celestial body, has an almost constant apparent scale, the system may misidentify red streetlights of certain angles and shapes as the sun, leading to a decrease in scene recognition accuracy. Stability verification can filter out the final interference such as streetlights. Through the serial logic of "four verifications", it is ensured that only when the camera continuously captures a round bright target with a position that matches the theoretical height of the sun and a stable scale during sunset, while the vehicle is in motion, is it identified as a "sunset flying car" scene.

[0053] Among them, the star scenes include the "Sunset Rollercoaster" scene and the "Sunrise Rollercoaster" scene corresponding to the sun, the "Chasing the Moon" scene corresponding to the moon, and the "Starry Sky Shuttle" scene corresponding to other stars. When the current star scene is the "Chasing the Moon" scene or the "Starry Sky Shuttle" scene, the feature in the image detection algorithm can be adjusted from "high-brightness circle" to features suitable for star points or the moon (such as the moon's crater contour recognition and bright star point detection), and the corresponding position prediction algorithm is also replaced accordingly.

[0054] Both the "Chasing the Moon" and "Starry Sky Travel" scenarios fall under the category of celestial scenarios, but they differ in the types of celestial targets and their spatiotemporal relationship with the vehicle. The "Chasing the Moon" scenario uses a single celestial target (the moon) as the identification object, and verifies it by accurately matching the theoretical spatial location information of the celestial target with environmental perception data. In contrast, the "Starry Sky Travel" scenario uses multiple celestial targets or the entire starry sky as the identification object, and verifies the distribution characteristics of star points and their overall characteristics that change over time to determine the starry sky-related driving scenario.

[0055] In this optional embodiment, in addition to spatial location matching verification, stability verification of the scale changes of candidate visual targets over a continuous time period is further introduced. This can take advantage of the objective characteristic that distant celestial bodies change scale slowly during vehicle movement to effectively eliminate nearby interfering objects, thereby further improving the robustness and anti-interference ability of celestial scene recognition.

[0056] Optionally, the stability verification of the scale changes of the candidate visual target over a continuous time period includes: S141: Track the candidate visual target in multiple consecutive image frames to obtain the scale parameters of the candidate visual target in each image frame.

[0057] Specifically, in combination Figure 4 As shown, candidate visual targets are tracked in multiple consecutive image frames to obtain the scale parameters of the candidate visual targets in each image frame. The scale parameters can be the radius, diameter, area, or equivalent scale parameters of the target region.

[0058] S142: Calculate the scale change rate of the candidate visual target over a continuous time period based on the scale parameter.

[0059] Specifically, the scale change rate of candidate visual targets over a continuous time period is calculated based on the scale parameter.

[0060] S143: In response to determining that the candidate visual target satisfies the scale stability condition based on the scale change rate, the stability verification is deemed successful.

[0061] Specifically, when the rate of scale change is less than or equal to a preset threshold, the candidate visual target is determined to meet the scale stability condition, thereby eliminating nearby interfering objects.

[0062] When tracking candidate visual targets, a KCF tracker can be used. This tracker can extract grayscale features and color name features within the region to balance tracking accuracy and computational cost. When tracking consecutive frames, Hough circle detection is no longer performed on the entire image. Instead, the initialized KCF tracker is directly called to quickly search and locate the target position near the target position in the previous frame and predict the new position and bounding box of the target in the current frame.

[0063] After each frame is successfully tracked, the equivalent pixel radius of the target bounding box in that frame is recorded. After the tracking sequence ends, the average radius and the deviation rate of each frame are calculated using the recorded radius values ​​of multiple frames. When the deviation rate is less than or equal to a preset threshold (e.g., 5%), it indicates that the target has a constant scale during the vehicle's movement, which is consistent with the distant characteristics of the sun. If the deviation rate fluctuates significantly (e.g., ≥45%), the target is determined to be a close-range object and is excluded.

[0064] In the tracking sequence, the target is allowed to be lost in a frame or have a confidence level below a threshold due to temporary occlusion (such as being blocked by trees or bridges).

[0065] In this optional embodiment, the candidate visual target is tracked and its scale change rate is calculated in multiple consecutive image frames. This allows for a quantitative judgment of the stability of the candidate visual target in a continuous time dimension, thereby avoiding misjudgment caused by instantaneous detection errors and improving the accuracy of scale stability verification and the stability of scene recognition results.

[0066] Optionally, the driving scenario includes a weather scenario, and determining the driving scenario based on the spatiotemporal data and the environmental perception data includes: S110B: Determine the meteorological triggering conditions corresponding to the meteorological scene based on the spatiotemporal data.

[0067] Specifically, in combination Figure 5 As shown, driving scenarios include weather scenarios (such as rainbow scenarios). The weather trigger conditions corresponding to the weather scenario are determined based on the vehicle's spatiotemporal data. For example, the vehicle's driving area is in a period after rain or meets specific weather conditions (the corresponding weather data is obtained based on the vehicle's location).

[0068] Among them, a meteorological scenario refers to a driving scenario related to a specific meteorological phenomenon determined during vehicle driving, based on the vehicle's time and geographical location information, combined with triggering factors related to meteorological conditions, and verified by environmental perception data to verify environmental characteristics related to meteorological phenomena. A meteorological scenario typically has one or more of the following technical characteristics: (1) the existence of meteorological triggering conditions: for example, time-related (such as after rain, a specific time period) or location-related (such as meteorological conditions prone to occur in a specific area); (2) the existence of perceptible meteorological environmental characteristics: for example, detectable through environmental perception data; exemplary meteorological scenarios include: (1) a rainbow scenario, which is a driving scenario determined during vehicle driving, based on meteorological triggering conditions such as after rain, and combined with environmental perception data to detect environmental visual information with rainbow characteristics; (2) a cloud change meteorological scenario, which is a driving scenario determined during vehicle driving. In the middle, the driving scene is determined based on specific meteorological conditions and the detection of obvious cloud structure or cloud movement characteristics by combining environmental perception data; (3) Special sky color meteorological scene, that is, the driving scene is determined based on changes in meteorological conditions and the detection of specific color distribution characteristics in the sky area by combining environmental perception data during vehicle driving; (4) Low visibility meteorological scene, that is, the driving scene is determined based on meteorological conditions and the detection of significantly reduced environmental visibility by combining environmental perception data during vehicle driving; It should be noted that the meteorological scene mainly focuses on the environmental performance caused by changes in meteorological conditions, and its identification depends on the joint judgment of meteorological triggering conditions and environmental perception data; Compared with the star scene with celestial targets as the core and the landscape scene with geographical regions as the core, the meteorological scene does not depend on the theoretical spatial location information of specific celestial bodies and the scene feature information of geographical landscapes.

[0069] Among them, meteorological triggering conditions refer to the conditions that limit the time range, spatial range, or environmental state of possible specific meteorological phenomena based on the spatiotemporal data of the vehicle during its driving process, and are used to trigger further environmental perception verification for the corresponding meteorological scenario. When the vehicle's spatiotemporal data meets the preset time and / or spatial conditions related to the probability of a certain meteorological phenomenon, it is considered to meet the corresponding meteorological triggering conditions. Taking the rainbow scenario as an example, the corresponding meteorological triggering conditions include: the vehicle's driving time is within a preset time range after the end of the rainfall, and the vehicle's position meets the lighting and spatial conditions that make rainbows likely to occur, such as the vehicle being in an open or unobstructed area within a certain time range after the rain; and cloud cover changes. Taking a weather scenario as an example, the corresponding weather triggering conditions include: the vehicle's travel time is during a specific weather change period, or the vehicle's location is in an area with significant cloud cover changes, such as a period when the weather changes from sunny to cloudy or from cloudy to sunny, or an area with significant terrain changes; taking a special sky color weather scenario as an example, the corresponding weather triggering conditions include: the vehicle's travel time is during a period when weather changes cause significant changes in sky color, such as a short period after a specific weather change, or an area with relatively stable lighting conditions; taking a low visibility weather scenario as an example, the corresponding weather triggering conditions include: the vehicle's travel time and location meet the conditions for the occurrence of low visibility weather phenomena, such as early morning or nighttime, or an area prone to fog formation.

[0070] S120B: When the vehicle's position meets the weather triggering conditions, candidate visual features corresponding to the weather scene are detected based on the environmental perception data.

[0071] Specifically, when the vehicle's current location meets the meteorological triggering conditions, candidate visual features corresponding to the meteorological scene are further detected based on environmental perception data, such as environmental image features with specific color distribution or morphological features.

[0072] Candidate visual features refer to visual attributes or image features extracted from environmental perception data to characterize the environment that may correspond to a certain driving scenario. Candidate visual features are used as the verification basis for determining whether the corresponding driving scenario exists. Since the environmental manifestations of different meteorological phenomena are completely different, different meteorological scenarios can correspond to different candidate visual features. For example, in a rainbow meteorological scenario, candidate visual features may include: arc-shaped structure features in the sky area, multi-color band distribution features, and spectral features with continuous gradient colors. For example, in a cloud change meteorological scenario, candidate visual features may include: texture features of clouds in the sky area, cloud block boundary features, and features of cloud distribution changing over time. For example, in a special sky color meteorological scenario, candidate visual features may include: overall color distribution features of the sky area and color change features that are significantly different from the normal sky color. For example, in a low visibility meteorological scenario, candidate visual features may include: features of reduced overall image contrast, features of reduced clarity of distant targets, or features of changes in image blur. In different meteorological scenarios, the candidate visual features used can be set according to the environmental manifestations of the corresponding meteorological phenomena. The candidate visual features corresponding to different meteorological scenarios may be the same or different, and are not limited to the examples listed in this specification.

[0073] S130B: When the candidate visual features meet the preset meteorological scene feature requirements, it is determined that the meteorological scene exists.

[0074] Specifically, when a candidate visual feature meets the preset meteorological scene feature requirements, it is determined that the meteorological scene exists in the current driving trip. That is, the candidate visual feature detected from the environmental perception data meets one or more of the following criteria: morphological features, color features, spatial distribution features, or change features. For example, in a rainbow meteorological scene, when the detected candidate visual feature includes an arc-shaped structure in the sky area, and the arc-shaped structure exhibits at least two continuous color distributions, and the color distribution remains stable across multiple consecutive image frames, the candidate visual feature is determined to meet the feature requirements of a rainbow meteorological scene. Similarly, in a cloud change meteorological scene, when the detected candidate visual feature includes cloud texture features in the sky area, and the cloud texture or cloud boundary remains stable across multiple consecutive image frames, the candidate visual feature meets the feature requirements of a rainbow meteorological scene. When significant changes occur within a time interval, the candidate visual feature is determined to meet the feature requirements of the cloud change meteorological scene. For example, in a special sky color meteorological scene, when the detected candidate visual feature indicates that the overall color distribution of the sky area deviates from the normal sky color range, and the color shift persists within a preset time period, the candidate visual feature is determined to meet the feature requirements of the special sky color meteorological scene. For example, in a low visibility meteorological scene, when the detected candidate visual feature indicates that the overall contrast or sharpness of the image is lower than a preset threshold, and this state remains consistent in multiple consecutive image frames, the candidate visual feature is determined to meet the feature requirements of the low visibility meteorological scene. It should be noted that the meteorological scene feature requirements corresponding to different meteorological scenes can be set according to actual applications, and the determination method is not limited to the above examples.

[0075] In this optional embodiment, the triggering conditions for the meteorological scene are set based on spatiotemporal data, and the candidate visual features are verified by combining environmental perception data when the triggering conditions are met. This enables joint judgment of meteorological data and visual information, avoiding misjudgments caused by relying solely on meteorological information or solely on image features, thereby improving the reliability and applicability of meteorological scene recognition.

[0076] Optionally, the driving scenario includes a landscape scenario, and determining the driving scenario based on the spatiotemporal data and the environmental perception data includes: S110C: Determine the spatial relationship between the vehicle's position and the preset landscape area based on the spatiotemporal data.

[0077] Specifically, in combination Figure 6 As shown, the driving scenario includes a landscape scenario. The spatial relationship between the vehicle's current position and the preset landscape area is determined based on the vehicle's spatiotemporal data. For example, whether the vehicle's driving route is close to the coastline, mountains, or areas with specific terrain features. Corresponding scenarios include coastal highway / mountain highway scenarios, etc.

[0078] Among them, the preset landscape area refers to the geographic area that is pre-marked based on geographic information data and associated with specific natural or cultural landscape features. It is used as a spatial reference range to determine whether a vehicle may be in the corresponding landscape scene. The preset landscape area can be set based on map data, geographic information system data or road attribute data, such as being marked based on coastline data, terrain slope data or road type data. The spatial relationship between the vehicle position and the preset landscape area refers to the spatial geometric relationship between the vehicle's current position information and the preset landscape area to determine whether the vehicle is located in, close to or traveling along the preset landscape area. Exemplary spatial relationships include: (1) inclusion relationship (located in): the vehicle's current position is within the boundary range of the preset landscape area, such as the vehicle being located in the marked coastal highway area; (2) distance relationship (adjacent): the distance between the vehicle's current position and the boundary of the preset landscape area is less than a preset distance threshold, such as the distance from the coastline is less than a certain distance; (3) direction consistency relationship (along the route): the vehicle's current position is within the boundary range of the preset landscape area, such as the distance from the coastline is less than a certain distance; (4) direction consistency relationship (along the route): the vehicle's current position is within the boundary range of the preset landscape area, such as the distance from the coastline is less than a certain distance; (5) direction consistency relationship (along the route): the vehicle's current position is within the boundary range of the preset landscape area, such as the distance from the coastline is less than a certain distance; (6) direction consistency relationship (along the route): the vehicle's current position is within the boundary range of the preset landscape area, such as the distance from the coastline is less than a certain distance; (7) direction consistency relationship (along the route): the vehicle's current position is within the boundary range of the preset landscape area, such as the distance from the coastline is less than a certain distance; (8) direction consistency relationship (along the route): the vehicle's current position is within the boundary range of the preset landscape area, such as the distance from the coastline is less than a certain distance; (9) direction consistency relationship ( The direction and the orientation of the preset landscape area are consistent within a certain angle range, for example, the vehicle travels along the coastline or ridge; (4) Path overlap relationship (path matching): the vehicle's travel route and the road segment marked as a landscape road partially or completely overlap on the map; different landscape scenes can correspond to different preset landscape areas. For example, the preset landscape area corresponding to the coastal highway scene can include: road areas near the coastline, road segments extending along the coastline, or road areas less than the preset threshold distance from the coastline. For example, the preset landscape area corresponding to the mountain road scene can include: road areas located in mountainous areas and whose road slope, curvature, or altitude change characteristics meet the preset conditions. For example, the preset landscape area corresponding to the canyon or high terrain landscape scene can include: areas with significant terrain undulations or roads extending along valleys. In the case where two landscape scenes have a certain degree of overlap, different landscape scenes can also correspond to the same preset landscape area. The specific settings can be based on the landscape scene type and are not limited to the examples listed in this manual.

[0079] S120C: Under the condition of satisfying the spatial relationship, detect candidate visual features corresponding to the landscape scene based on the environmental perception data.

[0080] Specifically, given the aforementioned spatial relationship, candidate visual features corresponding to the landscape scene are detected based on environmental perception data, such as sea level contours, continuous curve features, or terrain contour features.

[0081] Among them, satisfying the spatial relationship means that the spatial geometric relationship or path matching relationship between the vehicle's current position or driving route and the preset landscape area meets the spatial judgment conditions preset for the corresponding landscape scene. For example, in the coastal highway landscape scene, the spatial relationship satisfaction conditions may include: the vehicle's current position is located within the preset coastal road area, or the distance between the vehicle's current position and the coastline is less than a preset distance threshold, and the direction of the vehicle's driving route is consistent with the direction of the coastline within a preset angle range. For example, in the mountain road landscape scene, the spatial relationship satisfaction conditions may include: the vehicle's current position is located within the preset mountain road area, and the vehicle's driving route is presented within a continuous road segment. Significant curvature or slope changes are observed, such as in undulating terrain landscapes. Spatial relationship conditions may include: the vehicle's current location is within a predefined area of ​​significant terrain undulation, and the elevation change along the vehicle's route exceeds a predefined threshold. It should be noted that the candidate visual features may differ for different landscape scenarios. For example, coastal highway scenarios may focus on detecting sea level contour features, while mountain highway scenarios may focus on detecting continuous curves or terrain contour features. In special cases, they may also be partially the same. For different landscape scenarios, the specific criteria for determining spatial relationships can be set according to the landscape scenario type, and are not limited to the examples listed in this manual.

[0082] S130C: When the candidate visual features meet the preset landscape scene feature requirements, it is determined that the landscape scene exists.

[0083] Specifically, when the candidate visual features meet the preset landscape scene feature requirements, it is determined that there is a landscape scene in the current driving journey.

[0084] Among them, the preset landscape scene features refer to the visual feature conditions pre-set for different types of landscape scenes to characterize the typical manifestation of the landscape scene in environmental perception data. When one or more of the candidate visual features detected from the environmental perception data in terms of morphological features, spatial distribution features, or change features meet the preset landscape scene features of the corresponding landscape scene, the candidate visual features are determined to meet the preset landscape scene feature requirements. For example, the preset landscape scene features corresponding to a coastal highway landscape scene may include: the existence of continuously distributed horizontal contour features within the field of vision corresponding to the vehicle's driving direction, and the contour features remain stable in multiple image frames. When the detected candidate visual features include continuous horizontal edges located in the lower half of the image or in the distant area, and the edges remain stable in multiple consecutive image frames, it is determined that the coastal highway landscape scene feature requirements are met. For example, the coastal highway landscape scene features are determined to meet the preset landscape scene requirements. The preset landscape scene features corresponding to a scene may include: the road area exhibits a continuous curved distribution in the image, or the road edge line shows obvious directional changes in consecutive image frames. When the detected candidate visual features, including the road edge line, show multiple directional changes in consecutive image frames, and the magnitude of the changes exceeds a preset threshold, it is determined that the requirements for the mountain road landscape scene features are met. For example, the preset landscape scene features corresponding to a terrain undulation landscape scene may include: the horizon or terrain outline in the image shows obvious undulation changes. When the detected candidate visual features representing the height change of the terrain outline in the horizontal direction of the image exceed a preset threshold, it is determined that the requirements for the terrain undulation landscape scene features are met. It should be noted that the preset landscape scene features corresponding to different landscape scenes can be set according to the landscape scene type, and their specific forms and determination methods are not limited to the examples listed in this specification.

[0085] In this optional embodiment, landscape scene recognition is triggered based on the spatial relationship between the vehicle location and the preset landscape area, and candidate visual features are verified by combining environmental perception data. This can reduce the false detection probability of irrelevant areas while ensuring recognition efficiency, thereby improving the accuracy of landscape scene recognition and enhancing the adaptability of the method to different geographical environments.

[0086] Optionally, before displaying the trip review report corresponding to the driving scenario on a preset interface based on the multi-source data and the driving scenario, the method further includes: S310: The scene marker data corresponding to the driving scenario is associated and integrated with the basic trip data to generate scenario-based trip data.

[0087] Specifically, in combination Figure 7As shown, after the trip is completed, the scene marker data corresponding to the determined driving scenario is associated and integrated with the trip base data to generate scenario-based trip data. For example, the time interval corresponding to the scene marker data can be mapped to the driving trajectory and mileage information in the trip base data, thereby determining the position and duration of each driving scenario in the whole trip. The trip base data includes trip mileage information, driving route information and trip time information.

[0088] S320: Generate a trip review report corresponding to the driving scenario based on the scenario-based trip data.

[0089] Specifically, based on the scenario-based trip data, a trip review report corresponding to the driving scenario is generated, which is used to display scenario-based information of the driving trip to the user in a preset interface.

[0090] In this optional embodiment, after the trip is completed, the scene tag data corresponding to the driving scenario is associated and integrated with the basic trip data, and a trip review report is generated based on the scenario-based trip data. This can transform scattered trip data into structured scenario-based trip information, thereby improving the organization and readability of the trip review content.

[0091] Optionally, the trip review report corresponding to the driving scenario includes at least one of the following: scenario segment data, timeline data, scenario location information, and scenario summary information.

[0092] Specifically, the trip review report may include at least one of the following information: (1) Scene fragment data: continuous image frame data obtained from the driving time period corresponding to the driving scene. The continuous image frame data can be processed into short videos or image sequences to present the driving scene intuitively; (2) Time axis data: the start and end times of the driving scene are marked on the trip time axis to distinguish the driving scene occurrence interval from the normal driving interval; (3) Scene location information: based on the vehicle location information corresponding to the driving scene, the marked location of the driving scene on the map is generated to display the geographical location of the driving scene in the map interface; (4) Scene summary information: textual or graphic information used to summarize the driving scene.

[0093] In this optional embodiment, by introducing at least one of the following information in the trip review report: scene fragment data, timeline data, scene location information, and scene summary information corresponding to the driving scenario, the driving scenario can be expressed in a structured way from multiple dimensions such as visual presentation, time distribution, and spatial location. This makes the driving scenario no longer presented as a single text or simple markup, but as a multi-dimensional information closely related to the trip process, thereby improving the accuracy and comprehensibility of the trip review report in expressing the driving scenario.

[0094] Optionally, the scene summary information is generated based on the duration of the driving scene, trip feature information, or road type information.

[0095] Specifically, the scene summary information is generated based on at least one of the following: duration information, trip feature information, or road type information corresponding to the driving scene. For example, it can generate summary information reflecting the time characteristics of the driving scene based on the duration of the driving scene throughout the trip, or it can generate summary information reflecting the driving characteristics of the driving scene based on the type of road segment (such as highway, urban road, or mountain road) corresponding to the driving scene. The scene summary information can be generated in the form of natural language description to intuitively describe the characteristics of the driving scene to the user, such as "12-minute sunset drive on the national highway, golden afterglow accompanies you on your journey" to describe the time characteristics, spatial characteristics, or driving characteristics of the driving scene. It can also be displayed in the form of icons, tags, or combinations thereof.

[0096] In this optional embodiment, by generating scenario summary information based on the duration information, trip feature information, or road type information of the driving scenario, objective trip attributes related to the driving scenario can be transformed into summary content for summarizing the driving scenario. This allows the trip review report to effectively summarize the time characteristics, driving characteristics, or road attributes of the driving scenario without increasing the complexity of the presentation. It avoids the lack of adaptability caused by relying on manual editing or fixed templates to generate summary information. The scenario summary information can be dynamically generated according to different driving scenarios and different trip conditions, thereby improving the information density and expressive relevance of the trip review report.

[0097] The exemplary process for generating a trip review report is as follows: I. Itinerary and Scene Recognition: (1) Basic information of the trip (basic data of the trip): Starting point of the trip: City A; Ending point of the trip: City B; Total trip duration: 1 hour 20 minutes; Route of the trip: National Highway + Coastal Highway; Trip time: 18:10–19:30; (2) Driving scene recognition results (scene labeling data): Driving scene type: star scene (sunset flying car); scene start time: 18:32; scene end time: 18:44; scene duration: 12 minutes; scene corresponding location range: a continuous section of coastal highway; II. Generating Contextualized Trip Data: After the trip, the aforementioned contextualized data is linked and integrated with the basic trip data. For example, the time interval of "18:32–18:44" is mapped to the trip trajectory to determine the corresponding mileage interval, route location (coastal highway), and continuous image frame data within that time period. The resulting contextualized trip data includes: Contextualized: Sunset Flying Car; Duration: 12 minutes; Corresponding Route: National Highway G×× Coastal Section; Corresponding Track Point Set (Latitude and Longitude Sequence). III. Generate Trip Review Report: Based on the above scenario-based trip data, generate a trip review report corresponding to the driving scenario. The report may include the following: (1) Scene segment data (visual presentation): Select consecutive image frames from the forward-facing camera images collected between 18:32 and 18:44, and generate short video segments at a preset frame rate (e.g., 1fps). The generated scene segment data is used to show the visual effect of the sun gradually approaching the horizon during the vehicle's movement. (2) Timeline data (structured display of the trip): Mark the timeline of the trip: Regular travel period: 18:10–18:32; Sunset Rollercoaster Scene Time Range: 18:32–18:44; Regular operating hours: 18:44–19:30; (3) Scene location information (spatial dimension display): Map the latitude and longitude of the scene to the map interface and mark the route segment where the scene occurs on the map; (4) Scene summary information (general description): Generate summary information based on scene duration and road type, such as: "12-minute coastal highway sunset car"; IV. Presenting the Trip Review Report: Combined with Figure 10 As shown, the trip review report can be displayed in the following format on the in-vehicle terminal or mobile app: Top: Scene Name: Sunset Rollercoaster; Central area: Short video clips of scenes, map markers (coastal highway section); Bottom: Timeline (18:32–18:44), Scene summary information: "12-minute sunset roller coaster on the coastal highway".

[0098] like Figure 8 As shown in the figure, a driving scene recognition device 800 provided in this application embodiment includes: The first module 810 is used to acquire multi-source data related to the driving journey during vehicle operation, wherein the multi-source data includes the vehicle's spatiotemporal data and environmental perception data. The second module 820 is used to determine the driving scenario based on the spatiotemporal data and the environmental perception data; The third module 830 is used to display a trip review report corresponding to the driving scenario on a preset interface based on the multi-source data and the driving scenario.

[0099] like Figure 9 As shown in the embodiment of this application, an electronic device 900 includes a memory 920 and a processor 910; the memory 920 is used to store a computer program; the processor 910 is used to implement the driving scene recognition method as described above when the computer program is executed.

[0100] Alternatively, an electronic device 900 includes a memory 920 and a processor 910 coupled to the memory 920; the memory 920 is configured to store a computer program; and the processor 910 is configured to perform the following operations when the computer program is executed: During vehicle operation, multi-source data related to the driving journey is acquired, including the vehicle's spatiotemporal data and environmental perception data. The driving scenario is determined based on the spatiotemporal data and the environmental perception data. Based on the multi-source data and the driving scenario, a trip review report corresponding to the driving scenario is displayed on a preset interface.

[0101] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the driving scene recognition method described above.

[0102] Alternatively, a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the following operations: During vehicle operation, multi-source data related to the driving journey is acquired, including the vehicle's spatiotemporal data and environmental perception data. The driving scenario is determined based on the spatiotemporal data and the environmental perception data. Based on the multi-source data and the driving scenario, a trip review report corresponding to the driving scenario is displayed on a preset interface.

[0103] Electronic device 900, which can serve as a server or client in this application, is described below as an example of hardware devices that can be applied to various aspects of this application. Electronic device 900 is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic device 900 can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the application described and / or claimed herein.

[0104] Electronic device 900 includes a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or a computer program loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0105] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. In this application, 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 units can be selected to achieve the purpose of the embodiments of this application according to actual needs. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units can be implemented in hardware or as software functional units.

[0106] Although the above disclosure is provided, the scope of protection of this application is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of this application, and all such changes and modifications will fall within the scope of protection of this application.

Claims

1. A driving scene recognition method, characterized in that, include: During vehicle operation, multi-source data related to the driving journey is acquired, including the vehicle's spatiotemporal data and environmental perception data. The driving scenario is determined based on the spatiotemporal data and the environmental perception data. Based on the multi-source data and the driving scenario, a trip review report corresponding to the driving scenario is displayed on a preset interface.

2. The driving scene recognition method according to claim 1, characterized in that, The driving scenario includes a celestial scenario, and determining the driving scenario based on the spatiotemporal data and the environmental perception data includes: Based on the spatiotemporal data, determine the theoretical spatial position information of the celestial target corresponding to the celestial scene relative to the vehicle; Based on the environmental perception data, candidate visual targets corresponding to the celestial targets are detected; The spatial location information of the candidate visual target is matched and verified based on the theoretical spatial location information. When the matching verification is successful, it is determined that there is a star scene corresponding to the star target.

3. The driving scene recognition method according to claim 2, characterized in that, The theoretical spatial location information includes the elevation angle and azimuth angle of the celestial target, and the matching and verification of the spatial location information of the candidate visual target based on the theoretical spatial location information includes: The spatial location information of the candidate visual target is matched and verified with the theoretical spatial location information of the celestial target. When the difference between the elevation angle and azimuth angle of the candidate visual target and the elevation angle and azimuth angle of the celestial target is less than a preset angle threshold, the matching verification is deemed successful.

4. The driving scene recognition method according to claim 2, characterized in that, The step of determining the driving scenario based on the spatiotemporal data and the environmental perception data also includes: The stability of the scale changes of the candidate visual targets over a continuous time period is verified; Once the matching verification and stability verification are passed, it is determined that there is a star scene corresponding to the star target.

5. The driving scene recognition method according to claim 4, characterized in that, The stability verification of the scale changes of the candidate visual target over a continuous time period includes: The candidate visual target is tracked in multiple consecutive image frames to obtain the scale parameters of the candidate visual target in each image frame; The scale change rate of the candidate visual target over a continuous time period is calculated based on the scale parameter. In response to determining that the candidate visual target satisfies the scale stability condition based on the scale change rate, the stability verification is deemed successful.

6. The driving scene recognition method according to claim 1, characterized in that, Determining the driving scenario based on the spatiotemporal data and the environmental perception data includes: When the driving scenario includes a weather scenario, the weather triggering conditions corresponding to the weather scenario are determined based on the spatiotemporal data; when the vehicle's position meets the weather triggering conditions, candidate visual features corresponding to the weather scenario are detected based on the environmental perception data; when the candidate visual features meet the preset weather scenario feature requirements, it is determined that the weather scenario exists. And / or, When the driving scenario includes a landscape scenario, the spatial relationship between the vehicle's position and a preset landscape area is determined based on the spatiotemporal data; if the spatial relationship is satisfied, candidate visual features corresponding to the landscape scenario are detected based on the environmental perception data; when the candidate visual features meet the preset landscape scenario feature requirements, it is determined that the landscape scenario exists.

7. The driving scene recognition method according to claim 1, characterized in that, Before displaying the trip review report corresponding to the driving scenario on a preset interface based on the multi-source data and the driving scenario, the method further includes: The scenario-specific data corresponding to the driving scenario is linked and integrated with the basic trip data to generate scenario-specific trip data; A trip review report corresponding to the driving scenario is generated based on the scenario-based trip data.

8. The driving scene recognition method according to claim 1, characterized in that, The trip review report corresponding to the driving scenario includes at least one of the following: scenario segment data, timeline data, scenario location information, and scenario summary information.

9. An electronic device, characterized in that, Including memory and processor; The memory is used to store computer programs; The processor is configured to implement the driving scene recognition method as described in any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the driving scene recognition method as described in any one of claims 1 to 8.

Images