Collision detection method, apparatus and vehicle

Abstract

The application provides a collision detection method, device and vehicle, and the method comprises the following steps: determining an associated object of the vehicle from adjacent frame images of a driving area; predicting a collision time of the vehicle and the associated object based on a pixel width of the associated object in each frame image and a distance between the associated object and the vehicle in each frame image; and predicting a collision result of the vehicle and the associated object based on a motion trajectory of the vehicle within a target time length and a motion trajectory of the associated object within the target time length. The application can more accurately determine the motion trajectory change of the associated object by combining the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image, and can accurately predict the collision time of the vehicle and the associated object according to the motion trajectory change of the associated object, thereby accurately predicting the collision result of the vehicle and the associated object.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, and in particular to a collision detection method, device, and vehicle. Background Technology

[0002] In the automotive field, collision detection can be understood as detecting whether a collision will occur between two vehicles or between a vehicle and an obstacle. Currently, vehicle collision detection mainly relies on the driver's intuitive judgment; however, this intuition often deviates from the actual collision situation. In other words, the accuracy of a driver's prediction of collision detection based on their own intuition is relatively low. Summary of the Invention

[0003] This invention provides a collision detection method, device, and vehicle to address the deficiencies in the prior art.

[0004] This invention provides a collision detection method, comprising the following steps:

[0005] Identify associated objects of the vehicle from adjacent frame images of the driving area;

[0006] Based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image, the collision time between the vehicle and the associated object is predicted.

[0007] Based on the vehicle's trajectory within the target duration and the trajectory of the associated object within the target duration, the collision result between the vehicle and the associated object is predicted. The start time of the target duration is the acquisition time of the first frame image in the adjacent frame images, and the end time of the target duration is the collision time.

[0008] According to a collision detection method provided by the present invention, predicting the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image includes:

[0009] The first collision time is determined based on the pixel width of the associated object in each frame of the image and the acquisition time of each frame of the image.

[0010] The second collision time is determined based on the distance between the associated object and the vehicle in each frame of the image and the acquisition time of each frame of the image.

[0011] Based on the first collision time and the second collision time, the collision time between the vehicle and the associated object is determined.

[0012] According to a collision detection method provided by the present invention, determining the collision time between the vehicle and the associated object based on the first collision time and the second collision time includes:

[0013] Based on preset weights, the first collision time and the second collision time are weighted and added together to determine the collision time between the vehicle and the associated object.

[0014] According to a collision detection method provided by the present invention, the step of predicting the collision result between the vehicle and the associated object based on the vehicle's trajectory within a target time period and the trajectory of the associated object within the target time period includes:

[0015] If the trajectory of the vehicle within the target time period overlaps with the trajectory of the associated object within the target time period, then it is determined that there is a risk of collision between the vehicle and the associated object.

[0016] According to a collision detection method provided by the present invention, the step of predicting the collision result between the vehicle and the associated object further includes:

[0017] Obtain the actual collision result between the vehicle and the associated object;

[0018] If the actual collision result differs from the collision result, the vehicle braking system parameters are adjusted based on the actual collision result and the difference between the collision results.

[0019] According to a collision detection method provided by the present invention, obtaining the actual collision result between the vehicle and the associated object includes:

[0020] Based on the collision detection model, the actual collision result between the vehicle and the associated object is determined by applying the frame images of the driving area;

[0021] The collision detection model is trained based on sample frame images of the sample vehicle in the sample driving area, the sample associated objects of the sample vehicle, and the collision result labels between the sample vehicle and the sample associated objects.

[0022] The present invention also provides a collision detection device, comprising the following modules:

[0023] A determining unit is used to determine the associated objects of the vehicle from adjacent frame images of the driving area;

[0024] The prediction unit is used to predict the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image.

[0025] The detection unit is used to predict the collision result between the vehicle and the associated object based on the vehicle's trajectory within the target duration and the associated object's trajectory within the target duration. The start time of the target duration is the acquisition time of the start frame image in the adjacent frame images, and the end time of the target duration is the collision time.

[0026] The present invention also provides a vehicle, including: a collision detection device as described above and a vehicle braking system, wherein the collision detection device is used to adjust the parameters of the vehicle braking system based on the actual collision result and the difference between the collision result and the actual collision result when the actual collision result between the vehicle and the associated object is different from the collision result.

[0027] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the collision detection method as described above.

[0028] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the collision detection method as described above.

[0029] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the collision detection method as described above.

[0030] The collision detection method, apparatus, and vehicle provided by this invention capture the changes in the motion trajectory of related objects at a distance based on the pixel width of related objects in each frame of images, and capture the changes in the motion trajectory of related objects on a two-dimensional plane based on the distance between related objects and the vehicle in each frame of images. By combining the pixel width of related objects in each frame of images and the distance between related objects and the vehicle in each frame of images, the changes in the motion trajectory of related objects can be determined more accurately, thereby accurately predicting the collision time between the vehicle and related objects. Furthermore, based on the motion trajectory of the vehicle and the motion trajectory of related objects within a target time period, this invention can accurately determine whether the motion trajectories of the vehicle and related objects overlap within the target time period, thereby accurately predicting the collision result between the vehicle and related objects. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in this invention 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 invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0032] Figure 1 This is a schematic flowchart of the collision detection method provided by the present invention.

[0033] Figure 2 This is a flowchart illustrating an implementation of step 120 in the collision detection method provided by the present invention.

[0034] Figure 3 This is a schematic diagram of the OBB collision model provided by the present invention.

[0035] Figure 4 This is a schematic diagram of collision detection for each frame of the image provided by the present invention.

[0036] Figure 5 This is a schematic diagram of collision timing determination provided by the present invention.

[0037] Figure 6 This is a schematic diagram of the collision detection device provided by the present invention.

[0038] Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0040] Currently, vehicle collision detection primarily relies on the driver's intuitive judgment, meaning the driver intuitively predicts whether a collision will occur between two vehicles or between a vehicle and an obstacle based on a visually acquired image of the driving area. However, this intuition often deviates from the actual moment of collision. In other words, collision detection based on the driver's intuition has relatively low accuracy.

[0041] In response, this invention provides a collision detection method that uses adjacent frame images of the driving area as external input to objectively reflect the pixel width of the associated object and the distance between the vehicle and the associated object. Based on this objectively reflected pixel width and distance, the method predicts the collision time between the vehicle and the associated object and compares the movement trajectories of the vehicle and the associated object within a target time period to determine whether a collision will occur.

[0042] Figure 1 This is a flowchart illustrating the collision detection method provided by the present invention, as shown below. Figure 1As shown, the method includes steps 110, 120 and 130.

[0043] Step 110: Determine the associated objects of the vehicle from adjacent frame images of the driving area.

[0044] Specifically, the driving area can be understood as the range within which the vehicle travels, and adjacent frame images refer to two or more images within a continuous time interval. These adjacent frame images are used to represent road information, the size, shape, speed, and trajectory of related objects within the driving area. Optionally, a camera can be installed in front of the vehicle to collect adjacent frame images of the driving area while the vehicle is traveling within it; alternatively, occupants of the vehicle can collect adjacent frame images of the vehicle's movement using a camera device. This embodiment of the invention does not specifically limit the scope of these methods.

[0045] It is understandable that after acquiring adjacent frame images, image preprocessing can be performed on the adjacent frame images, such as grayscale conversion, noise reduction, image enhancement, geometric transformation, etc., to obtain high-quality adjacent frame images, thereby enabling accurate collision detection based on high-quality adjacent frame images.

[0046] Furthermore, associated objects refer to objects that pose a potential collision risk to the vehicle during its movement. These objects can be other vehicles, pedestrians, non-motorized vehicles, road obstacles, etc. After acquiring adjacent frame images of the driving area, image recognition can be performed on these adjacent frame images to identify multiple candidate objects. Based on the positional changes of each candidate object in the adjacent frame images, trajectory tracking is performed to determine the motion trajectory of each candidate object. Based on the motion trajectory of each candidate object, the relative positional relationship between each candidate object and the vehicle is determined. If any candidate object is in the same lane as the vehicle and is relatively close, it is considered that the candidate object may pose a collision risk to the vehicle, and in this case, the candidate object can be designated as an associated object.

[0047] In addition, the associated physical objects can be determined by multi-frame tracking, which involves performing target detection on adjacent frames to obtain the target objects in each frame and calculating the similarity between the target objects in each frame. This allows objects with high similarity in each frame to be associated, and these objects with high similarity are the associated objects.

[0048] Step 120: Based on the pixel width of the associated objects in each frame image and the distance between the associated objects and the vehicle in each frame image, predict the collision time between the vehicle and the associated objects.

[0049] Specifically, the pixel width of an associated object refers to the number of horizontal pixels occupied by the associated object in each frame of the image. Pixel width can be used to characterize the actual width of the associated object. For example, the actual width of the associated object can be determined based on the following formula:

[0050] ;

[0051] in, Indicates the actual width of the associated object. Indicates the pixel width of the associated object. This indicates the actual distance between the associated object and the vehicle. This indicates the focal length of the camera.

[0052] Although distant objects may have a small number of pixels in an image, even a tiny change in pixel width can indicate significant motion in reality. This invention, based on the pixel width of distant objects, can precisely detect subtle changes in them, thereby enabling early detection of potential collisions between these objects and vehicles.

[0053] However, the pixel width change of the associated object cannot accurately represent the complex motion of the object on the two-dimensional plane. For example, when the associated object rotates in the two-dimensional plane, this change will not be reflected in the pixel width. Therefore, the change of the trajectory of the associated object on the two-dimensional plane cannot be known through the pixel width.

[0054] Unlike pixel width, the distance between the associated object and the vehicle in each frame of the image takes into account the positional information of the associated object and the vehicle on the two-dimensional plane. Therefore, the relative positional relationship between the associated object and the vehicle on the two-dimensional plane can be determined based on their positions. The change in the relative positional relationship can determine the motion trajectory of the associated object, which reflects the change in the motion trajectory of the associated object on the two-dimensional plane.

[0055] However, the distance between the associated object and the vehicle in each frame focuses on the motion trajectory on a two-dimensional plane, which may introduce measurement errors for changes in the motion of distant associated objects. For example, when the motion of a distant associated object changes, the change in distance between the associated object and the vehicle in each frame may be very small, thus failing to reflect the motion changes of the distant associated object and consequently failing to capture potential collision hazards of distant associated objects in a timely manner.

[0056] In summary, the embodiments of the present invention capture the changes in the motion trajectory of related objects at a distance based on the pixel width of related objects in each frame of images, and capture the changes in the motion trajectory of related objects on a two-dimensional plane based on the distance between related objects and vehicles in each frame of images. Thus, by combining the pixel width of related objects in each frame of images and the distance between related objects and vehicles in each frame of images, the changes in the motion trajectory of related objects can be determined more accurately, and the collision time between the vehicle and related objects can be accurately predicted based on the changes in the motion trajectory of related objects.

[0057] For example, the actual distance between the vehicle and the associated object and the relative speed between the vehicle and the associated object can be determined by combining the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image, and the collision time between the vehicle and the associated object can be determined based on the actual distance and the relative speed.

[0058] Step 130: Based on the vehicle's trajectory and the trajectory of the associated object within the target duration, predict the collision result between the vehicle and the associated object. The start time of the target duration is the acquisition time of the first frame image in the adjacent frame image, and the end time of the target duration is the collision time.

[0059] Specifically, the predicted collision time refers to the moment when a collision is likely to occur between the vehicle and the associated object as they continue moving in their current state of motion. The reason it is described as a potential collision moment is that the collision time is determined based on relative speed and relative distance. In other words, the collision indicator suggests the moment when a collision is likely between the vehicle and the associated object at the current relative speed and distance. However, a collision requires that the trajectories of the vehicle and the associated object overlap. If the vehicle merely catches up with the associated object in parallel, but their trajectories do not overlap, then no collision occurs.

[0060] In this embodiment of the invention, after determining the collision time, the acquisition time of the initial frame image is used as the start time of the target duration, and the collision time is used as the end time of the target duration. The trajectory of the vehicle and the trajectory of the associated object within the target duration are then determined. If the two trajectories overlap, it indicates that there is a collision risk between the vehicle and the associated object within the target duration. If the two trajectories do not overlap, it indicates that there is no collision risk between the vehicle and the associated object within the target duration.

[0061] The collision detection method provided in this invention captures the changes in the motion trajectory of related objects at a distance based on the pixel width of related objects in each frame of images, and captures the changes in the motion trajectory of related objects on a two-dimensional plane based on the distance between related objects and vehicles in each frame of images. By combining the pixel width of related objects in each frame of images and the distance between related objects and vehicles in each frame of images, the changes in the motion trajectory of related objects can be determined more accurately, thereby accurately predicting the collision time between the vehicle and related objects. Furthermore, based on the motion trajectory of the vehicle and the motion trajectory of related objects within a target time period, this invention can accurately determine whether the motion trajectories of the vehicle and related objects overlap within the target time period, thereby accurately predicting the collision result between the vehicle and related objects.

[0062] Based on the above embodiments, Figure 2 This is a flowchart illustrating an implementation method for step 120 of the collision detection method provided by the present invention, as shown below. Figure 2 As shown, step 120 predicts the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image, including:

[0063] Step 121: Determine the first collision time based on the pixel width of the associated object in each frame image and the acquisition time of each frame image.

[0064] Specifically, the acquisition time of each frame can be understood as the timestamp corresponding to each frame. For example, the first collision time can be determined based on the pixel width of the associated object in two adjacent frames and the acquisition time of the two adjacent frames. Alternatively, multiple sets of two adjacent frames can be selected, and the first collision time corresponding to each set of two adjacent frames can be calculated. The average of the first collision times of each set can then be used as the final first collision time.

[0065] Taking the calculation of the first collision time between two adjacent frames as an example, the first collision time can be determined based on the following formula:

[0066] ;

[0067] in, Indicates the moment of the first collision. Indicates the acquisition time of the starting frame. This indicates the acquisition time of the next frame corresponding to the start frame. This represents the pixel width of the associated object in the starting frame. This indicates the pixel width of the associated object in the next frame corresponding to the starting frame.

[0068] Optionally, before determining the moment of the first collision, each frame of the image can be projected into a perspective view, so that the changes in the motion trajectory of related objects in the distance can be observed in detail.

[0069] Step 122: Determine the second collision time based on the distance between the associated object and the vehicle in each frame of the image and the acquisition time of each frame of the image.

[0070] Specifically, the acquisition time of each frame can be understood as the timestamp corresponding to each frame. For example, the second collision time can be determined based on the distance between the associated object and the vehicle in two adjacent frames, and the acquisition time of those two adjacent frames. Alternatively, multiple sets of adjacent frames can be selected, and the second collision time corresponding to each set can be calculated. The average of the second collision times from each set can then be used as the final second collision time.

[0071] Taking the calculation of the second collision time between two adjacent frames as an example, the second collision time can be determined based on the following formula:

[0072] ;

[0073] in, Indicates the moment of the second collision. Indicates the acquisition time of the starting frame. This indicates the acquisition time of the next frame corresponding to the start frame. This indicates the distance between the associated object and the vehicle in the initial frame. This indicates the distance between the associated object and the vehicle in the next frame corresponding to the starting frame.

[0074] Optionally, in order to improve the perception, understanding and prediction of the surrounding environment, before determining the second collision moment, the adjacent frame images can be transformed into a bird's eye view (BEV) transformation, that is, the adjacent frame images can be transformed into a view centered on the vehicle and looking down at the vehicle's surrounding environment from a high altitude.

[0075] Step 123: Based on the first collision time and the second collision time, determine the collision time between the vehicle and the associated object.

[0076] Specifically, the first collision time is determined based on the pixel width of the associated object in each frame of the image, thus accurately representing the collision risk of the distant associated object. The second collision time is determined based on the distance between the associated object and the vehicle in each frame of the image, thus accurately representing the collision risk of the associated object in the two-dimensional plane.

[0077] In this regard, the embodiments of the present invention combine the first collision time and the second collision time to determine the collision time between the vehicle and the associated object, taking into account both the collision hazard of the distant associated object and the collision hazard of the associated object on the two-dimensional plane.

[0078] Optionally, the nearest collision time can be selected from the first collision time and the second collision time as the collision time between the vehicle and the associated object, or the average of the first collision time and the second collision time can be used as the collision time between the vehicle and the associated object. This embodiment of the invention does not specifically limit this.

[0079] Based on any of the above embodiments, determining the collision time between the vehicle and the associated object based on the first collision time and the second collision time includes:

[0080] Based on preset weights, the first collision moment and the second collision moment are weighted and added together to determine the collision moment between the vehicle and the associated object.

[0081] Specifically, while the first collision moment can characterize the collision hazard of a distant associated object, it cannot characterize the collision hazard of that associated object in a two-dimensional plane. Similarly, while the second collision moment can characterize the collision hazard of an associated object in a two-dimensional plane, it cannot reflect the collision hazard of a distant associated object. Therefore, this embodiment of the invention, based on preset weights, adds the first and second collision moments together, so that the determined collision moment reflects both the collision hazard of a distant associated object and the collision hazard of that associated object in a two-dimensional plane.

[0082] Alternatively, the moment of collision between the vehicle and the associated object can be determined based on the following formula:

[0083] ;

[0084] in, Indicates the moment of collision between the vehicle and the associated object. Indicates the moment of the first collision. Indicates the moment of the second collision. This indicates the weight corresponding to the first collision moment. This represents the weight corresponding to the second collision moment, where, .

[0085] Based on any of the above embodiments, based on the vehicle's trajectory within a target time period and the trajectory of the associated object within the target time period, the collision result between the vehicle and the associated object is predicted, including:

[0086] If the vehicle's trajectory overlaps with that of a related object within the target time period, then a collision risk is identified between the vehicle and the related object.

[0087] Specifically, the overlap here can be understood as the vehicle's position being the same as the position of a related object at any given moment within the target time period. If there is any moment within the target time period where the vehicle's position is the same as the position of a related object, it indicates that there is a risk of collision between the vehicle and the related object. Conversely, if there is no moment within the target time period where the vehicle's position is the same as the position of a related object, it indicates that there is no risk of collision between the vehicle and the related object.

[0088] Figure 3 This is a schematic diagram of the OBB collision model provided by the present invention, as shown below. Figure 3 As shown, OBB (Oriented Bounding Box) is a bounding box built based on the shape and orientation of an object. It determines the corresponding OBBs of the vehicle and the target object within the target time period, and projects the two OBBs onto the two axes of the coordinate system. It then determines whether the projected OBBs on the two axes overlap. If they do, it indicates a collision risk between the vehicle and the associated object; otherwise, it indicates no collision risk. According to... Figure 3 As can be seen, the OBB projections on both axes overlap, indicating a collision risk between the vehicle and related objects.

[0089] Figure 4 This is a schematic diagram of collision detection for each frame of the image provided by the present invention, such as... Figure 4 As shown, within the target duration, corresponding frame images are collected every 0.1s, and each frame image is traversed to determine whether there is overlap between the vehicle and the corresponding OBB of the associated object in each frame image. If so, it is determined that there is a collision; otherwise, it is determined that there is no collision.

[0090] Furthermore, with the development of autonomous driving technology, the Automatic Emergency Braking (AEB) system, as a crucial safety feature, is vital for ensuring driving safety through its accuracy and reliability. However, AEB may be falsely triggered in certain situations, leading to unnecessary emergency braking, affecting the driving experience, and even causing accidents. The reason for false AEB triggering lies in the miscalculation of the collision outcome, meaning the predicted collision result differs from the actual collision result.

[0091] Therefore, it is necessary to evaluate the AEB system to confirm whether false triggering occurs. Currently, AEB system evaluation mainly relies on real-vehicle testing, but this method is costly, inefficient, and cannot cover all possible false triggering scenarios. In addition, real-vehicle testing often cannot fully simulate the complex environment of false triggering, leading to limitations in the evaluation results.

[0092] To address this, this embodiment of the invention uses the method described above to determine the collision result and compares it with the actual collision result to assess whether the AEB system might be falsely triggered. This method eliminates the need for real-vehicle testing, significantly reducing testing costs. Specifically, the evaluation method is as follows:

[0093] Predicting the collision outcome between the vehicle and related objects, and then including:

[0094] Obtain the actual collision results between the vehicle and related objects;

[0095] If the actual collision result differs from the actual collision result, the vehicle braking system parameters will be adjusted based on the actual collision result and the difference between the collision results.

[0096] Specifically, the actual collision result refers to the real outcome of whether a vehicle collides with an associated object. This actual collision result can be obtained through simulation testing or through monitoring by sensors (such as radar) installed on the vehicle.

[0097] If the actual collision result differs from the expected collision result, it indicates that the vehicle's braking system (such as the AEB system) may have been falsely triggered. In this case, the actual collision result and the differences between the results can be analyzed to identify the parameters that need adjustment, and the vehicle's braking system parameters can be adjusted accordingly. These adjusted parameters may include brake force distribution, braking response time, and braking distance.

[0098] Therefore, the embodiments of the present invention compare the actual collision results with the collision results, and adjust the vehicle braking system parameters when the two are inconsistent, so as to avoid the vehicle braking system from being falsely triggered.

[0099] Based on any of the above embodiments, obtaining the actual collision result between the vehicle and the associated object includes:

[0100] Based on the collision detection model, frame images of the driving area are used to determine the actual collision results between the vehicle and related objects;

[0101] The collision detection model is trained based on sample frame images of sample vehicles in sample driving areas, sample associated objects of sample vehicles, and collision result labels between sample vehicles and sample associated objects.

[0102] Specifically, sample frame images of the sample vehicle in the sample driving area can be obtained from open-source datasets, such as CNCAP2024, and the scenarios involved can include complex environments such as urban roads, highways, and rainy / snowy weather. In selecting sample frame images, frames corresponding to manual emergency braking can be removed to avoid interference with subsequent detection results.

[0103] After acquiring sample frame images, the images are annotated, including the annotation of obstacles and lane lines. Obstacle annotations include objects that may interact with the vehicle, such as vehicles, pedestrians, and bicycles. Lane line annotations are used to determine the vehicle's travel path and lane boundaries.

[0104] In addition to the above annotations, the collision results between the sample vehicle and the associated object also need to be annotated, i.e., the collision result label. This label can be the collision time between the sample vehicle and the associated object, which can be determined based on the relative distance and relative speed between the sample vehicle and the associated object.

[0105] In addition, before training based on the sample frame images of the sample vehicle in the sample driving area, the sample associated objects of the sample vehicle, and the collision results between the sample vehicle and the sample associated objects, a region of interest (FOI) can be labeled in the sample frame images. The FOI usually includes an area within a certain distance and angle in front of the sample vehicle, so that the sample associated objects can be identified by focusing on this area.

[0106] After training the collision detection model, the actual collision results between the vehicle and related objects can be determined by applying frame images of the driving area based on the collision detection model.

[0107] Based on any of the above embodiments, the present invention also provides a collision detection method, comprising:

[0108] Figure 5 This is a schematic diagram of collision timing determination provided by the present invention, as shown below. Figure 5 As shown, the vehicle's associated objects are determined from adjacent frames of the driving area using a multi-frame tracking method. From a perspective viewpoint, the first collision time is determined based on the pixel widths (widthT0 and widthT1) of the associated objects in each frame, and the acquisition times (T0 and T1) of each frame. From a BEV perspective, the second collision time is determined based on the distances (DepthT0 and DepthT1) between the associated objects and the vehicle in each frame, and the acquisition times of each frame. Then, based on preset weights, the first and second collision times are weighted and summed to determine the collision time between the vehicle and the associated objects.

[0109] Within the target duration, corresponding frame images are acquired every 0.1 seconds. Each frame image is then traversed, and it is determined whether the OBBs corresponding to the vehicle and the associated object overlap in each frame image. If they do, a collision is determined; otherwise, no collision is determined. Finally, the collision result between the vehicle and the associated object is obtained. The start time of the target duration is the acquisition time of the first frame image in the adjacent frame images, and the end time of the target duration is the collision time.

[0110] Next, based on the collision detection model, the actual collision results between the vehicle and the associated objects are determined by applying the frame images of the driving area. The collision detection model is trained based on the sample frame images of the sample vehicle in the sample driving area, the sample associated objects of the sample vehicle, and the collision result labels between the sample vehicle and the sample associated objects.

[0111] If the actual collision result differs from the actual collision result, the vehicle braking system parameters will be adjusted based on the actual collision result and the difference between the collision results.

[0112] The collision detection device provided by the present invention is described below. The collision detection device described below and the collision detection method described above can be referred to in correspondence.

[0113] Based on any of the above embodiments Figure 6 This is a schematic diagram of the collision detection device provided by the present invention, as shown below. Figure 6 As shown, the device includes:

[0114] The determining unit 610 is used to determine the associated objects of the vehicle from adjacent frame images of the driving area;

[0115] The prediction unit 620 is used to predict the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image.

[0116] The detection unit 630 is used to predict the collision result between the vehicle and the associated object based on the vehicle's trajectory within the target duration and the trajectory of the associated object within the target duration. The start time of the target duration is the acquisition time of the first frame image in the adjacent frame image, and the end time of the target duration is the collision time.

[0117] The collision detection device provided in this invention captures the changes in the motion trajectory of related objects at a distance based on the pixel width of related objects in each frame of images, and captures the changes in the motion trajectory of related objects on a two-dimensional plane based on the distance between related objects and vehicles in each frame of images. By combining the pixel width of related objects in each frame of images and the distance between related objects and vehicles in each frame of images, the changes in the motion trajectory of related objects can be determined more accurately, thereby accurately predicting the collision time between the vehicle and related objects based on the changes in the motion trajectory of related objects. Furthermore, based on the motion trajectory of the vehicle and the motion trajectory of related objects within a target time period, this invention can accurately determine whether the motion trajectories of the vehicle and related objects overlap within the target time period, thereby accurately predicting the collision result between the vehicle and related objects.

[0118] Based on any of the above embodiments, based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image, the collision time between the vehicle and the associated object is predicted, including:

[0119] The first collision time is determined based on the pixel width of the associated objects in each frame and the acquisition time of each frame.

[0120] The second collision time is determined based on the distance between the associated object and the vehicle in each frame of the image, as well as the acquisition time of each frame of the image.

[0121] Based on the first and second collision times, the collision times between the vehicle and the associated objects are determined.

[0122] Based on any of the above embodiments, determining the collision time between the vehicle and the associated object based on the first collision time and the second collision time includes:

[0123] Based on preset weights, the first collision moment and the second collision moment are weighted and added together to determine the collision moment between the vehicle and the associated object.

[0124] Based on any of the above embodiments, based on the vehicle's trajectory within a target time period and the trajectory of the associated object within the target time period, the collision result between the vehicle and the associated object is predicted, including:

[0125] If the vehicle's trajectory overlaps with that of a related object within the target time period, then a collision risk is identified between the vehicle and the related object.

[0126] Based on any of the above embodiments, the collision result between the vehicle and the associated object is predicted, and then the process further includes:

[0127] Obtain the actual collision results between the vehicle and related objects;

[0128] If the actual collision result differs from the actual collision result, the vehicle braking system parameters will be adjusted based on the actual collision result and the difference between the collision results.

[0129] Based on any of the above embodiments, obtaining the actual collision result between the vehicle and the associated object includes:

[0130] Based on the collision detection model, frame images of the driving area are used to determine the actual collision results between the vehicle and related objects;

[0131] The collision detection model is trained based on sample frame images of sample vehicles in sample driving areas, sample associated objects of sample vehicles, and collision result labels between sample vehicles and sample associated objects.

[0132] Based on any of the above embodiments, the present invention also provides a vehicle, including: a collision detection device as described in any of the above embodiments and a vehicle braking system, wherein the collision detection device is used to adjust the parameters of the vehicle braking system based on the actual collision result and the difference between the collision result and the actual collision result when the collision result between the vehicle and the associated object is different.

[0133] Specifically, if the actual collision result differs from the expected collision result, it indicates that the vehicle's braking system (such as the AEB system) may have been falsely triggered. In this case, the actual collision result and the differences between the results can be analyzed to identify the parameters that need adjustment, and the vehicle's braking system parameters can be adjusted accordingly. These adjusted parameters may include brake force distribution, braking response time, and braking distance.

[0134] Therefore, the embodiments of the present invention compare the actual collision results with the collision results, and adjust the vehicle braking system parameters when the two are inconsistent, so as to avoid the vehicle braking system from being falsely triggered.

[0135] Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 7 As shown, the electronic device may include a processor 710, a communications interface 720, a memory 730, and a communication bus 740, wherein the processor 710, communications interface 720, and memory 730 communicate with each other via the communication bus 740. The processor 710 can call logical instructions in the memory 730 to execute a collision detection method, which includes: determining associated objects of the vehicle from adjacent frame images of the driving area; predicting the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image; and predicting the collision result between the vehicle and the associated object based on the vehicle's trajectory within a target duration and the trajectory of the associated object within the target duration, wherein the start time of the target duration is the acquisition time of the first frame image in the adjacent frame images, and the end time of the target duration is the collision time.

[0136] Furthermore, the logical instructions in the aforementioned memory 730 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, essentially, 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.

[0137] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the collision detection method provided by the above methods. The method includes: determining the associated object of the vehicle from adjacent frame images of the driving area; predicting the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image; and predicting the collision result between the vehicle and the associated object based on the movement trajectory of the vehicle within a target time period and the movement trajectory of the associated object within the target time period, wherein the start time of the target time period is the acquisition time of the start frame image in the adjacent frame images, and the end time of the target time period is the collision time.

[0138] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the collision detection method provided by the methods described above. The method includes: determining associated objects of a vehicle from adjacent frame images of a driving area; predicting the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image; and predicting the collision result between the vehicle and the associated object based on the vehicle's trajectory within a target duration and the trajectory of the associated object within the target duration, wherein the start time of the target duration is the acquisition time of the starting frame image in the adjacent frame images, and the end time of the target duration is the collision time.

[0139] 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.

[0140] 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.

[0141] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A collision detection method, characterized in that, include: Identify associated objects of the vehicle from adjacent frame images of the driving area; Based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image, the collision time between the vehicle and the associated object is predicted. Based on the vehicle's trajectory within the target duration and the trajectory of the associated object within the target duration, the collision result between the vehicle and the associated object is predicted. The start time of the target duration is the acquisition time of the start frame image in the adjacent frame images, and the end time of the target duration is the collision time. The method of predicting the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image includes: The first collision time is determined based on the pixel width of the associated object in each frame of the image and the acquisition time of each frame of the image. The second collision time is determined based on the distance between the associated object and the vehicle in each frame of the image and the acquisition time of each frame of the image. Based on the first collision time and the second collision time, the collision time between the vehicle and the associated object is determined.

2. The collision detection method according to claim 1, characterized in that, Determining the collision time between the vehicle and the associated object based on the first collision time and the second collision time includes: Based on preset weights, the first collision time and the second collision time are weighted and added together to determine the collision time between the vehicle and the associated object.

3. The collision detection method according to any one of claims 1 to 2, characterized in that, The method of predicting the collision result between the vehicle and the associated object based on the vehicle's trajectory within the target time period and the trajectory of the associated object within the target time period includes: If the trajectory of the vehicle within the target time period overlaps with the trajectory of the associated object within the target time period, then it is determined that there is a risk of collision between the vehicle and the associated object.

4. The collision detection method according to any one of claims 1 to 2, characterized in that, The method of predicting the collision result between the vehicle and the associated object further includes: Obtain the actual collision result between the vehicle and the associated object; If the actual collision result differs from the collision result, the vehicle braking system parameters are adjusted based on the actual collision result and the difference between the collision results.

5. The collision detection method according to claim 4, characterized in that, The step of obtaining the actual collision result between the vehicle and the associated object includes: Based on the collision detection model, the actual collision result between the vehicle and the associated object is determined by applying the frame images of the driving area; The collision detection model is trained based on sample frame images of the sample vehicle in the sample driving area, the sample associated objects of the sample vehicle, and the collision result labels between the sample vehicle and the sample associated objects.

6. A collision detection device, characterized in that, include: A determining unit is used to determine the associated objects of the vehicle from adjacent frame images of the driving area; The prediction unit is used to predict the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image. The detection unit is used to predict the collision result between the vehicle and the associated object based on the vehicle's trajectory within the target duration and the associated object's trajectory within the target duration. The start time of the target duration is the acquisition time of the start frame image in the adjacent frame images, and the end time of the target duration is the collision time. The method of predicting the collision time between the vehicle and the associated object based on the pixel width of the associated object in each frame image and the distance between the associated object and the vehicle in each frame image includes: The first collision time is determined based on the pixel width of the associated object in each frame of the image and the acquisition time of each frame of the image. The second collision time is determined based on the distance between the associated object and the vehicle in each frame of the image and the acquisition time of each frame of the image. Based on the first collision time and the second collision time, the collision time between the vehicle and the associated object is determined.

7. A vehicle, characterized in that, include: The collision detection device and vehicle braking system as described in claim 6, wherein the collision detection device is used to adjust the parameters of the vehicle braking system based on the actual collision result and the difference between the collision result and the actual collision result when the actual collision result between the vehicle and the associated object is different from the collision result.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the collision detection method as described in any one of claims 1 to 5.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the collision detection method as described in any one of claims 1 to 5.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the collision detection method as described in any one of claims 1 to 5.

Images