Vehicle parking processing method, device and equipment based on security vehicle management terminal

Abstract

This disclosure presents a method, apparatus, and device for handling vehicle parking based on a security vehicle management terminal. One specific implementation of the method includes: determining at least one vehicle information; for each vehicle information, performing the following sending steps: determining whether the vehicle information is target vehicle information; in response to determining that it is target vehicle information, acquiring the vehicle parking location; determining at least one camera device information; acquiring a multi-directional video set in real time; generating parking hazard information and target object intent information based on the multi-directional video set; in response to determining that the value corresponding to the parking hazard information is within a first interval, sending the parking hazard information and the intent information to an in-vehicle central control device, and sending a linked SMS message to a user terminal. This implementation is related to the security field and can achieve real-time and accurate security monitoring of parked vehicles, effectively and promptly preventing dangers to parked vehicles.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of computer technology, and more specifically to a vehicle parking processing method, apparatus, and device based on a security vehicle management terminal. Background Technology

[0002] Currently, with the continuous development of artificial intelligence, the related technologies of smart parking lots are becoming increasingly mature, and the security management of smart parking lots is becoming increasingly important. The security monitoring of parked vehicles in parking lots typically employs a method where the vehicle's built-in security devices utilize artificial intelligence algorithms to ensure vehicle safety.

[0003] However, when using the above methods to ensure the safety of parked vehicles, the following technical problems often arise:

[0004] First, relying on the onboard safety devices of parked vehicles to ensure their safety in real time requires constant power supply to these devices, resulting in a waste of resources. Furthermore, the limited safety distance monitored by parked vehicles leaves security loopholes, creating potential safety hazards.

[0005] Secondly, the accuracy of intent recognition directly impacts vehicle safety, and the accuracy of intent recognition is an aspect that needs to be accurately guaranteed.

[0006] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0008] Some embodiments of this disclosure propose a vehicle parking processing method, apparatus, and equipment based on a security vehicle management terminal to solve one or more of the technical problems mentioned in the background section above.

[0009] In a first aspect, some embodiments of this disclosure provide a vehicle parking management method, including: using a set of cameras installed at a target parking area to determine at least one vehicle information corresponding to at least one parked vehicle; for each of the at least one vehicle information, performing the following transmission steps: determining whether the vehicle information is target vehicle information, wherein the target vehicle information is vehicle information that has been registered and authorized for parking security control; in response to determining that it is target vehicle information, obtaining the vehicle parking location corresponding to the vehicle information; determining at least one vehicle from the set of cameras adjacent to the vehicle parking location. Information on at least one camera device corresponding to a camera device; real-time acquisition of a multi-directional video set of the vehicle information based on the at least one camera device information; generation of parking hazard information and target object intent information based on the multi-directional video set, wherein the target object is an object moving toward the parking location of the vehicle; in response to determining that the value corresponding to the parking hazard information is in a first interval, sending the parking hazard information and the intent information to the in-vehicle central control device corresponding to the vehicle information, and sending a link SMS message for the multi-directional video set and the intent information to the user terminal of the registered object corresponding to the vehicle information.

[0010] Secondly, some embodiments of this disclosure provide a vehicle parking processing device, including: a determining unit configured to determine at least one vehicle information corresponding to at least one parked vehicle using a set of cameras installed in a target parking area; and an execution unit configured to perform the following sending steps for each of the at least one vehicle information: determining whether the vehicle information is target vehicle information, wherein the target vehicle information is vehicle information that has been registered and authorized for parking security control; in response to determining that it is target vehicle information, obtaining the vehicle parking location corresponding to the vehicle information; and determining the vehicle parking location adjacent to the vehicle parking location. At least one camera device information corresponding to at least one camera device in the camera device set; based on the at least one camera device information, a multi-directional video set of the vehicle information is acquired in real time; based on the multi-directional video set, parking hazard information and target object intent information are generated, wherein the target object is an object moving toward the parking position of the vehicle; in response to determining that the value corresponding to the parking hazard information is in a first interval, the parking hazard information and the intent information are sent to the in-vehicle central control device corresponding to the vehicle information, and a link SMS message for the multi-directional video set and the intent information is sent to the user terminal of the registered object corresponding to the vehicle information.

[0011] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, such that when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0012] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0013] The various embodiments of this disclosure have the following beneficial effects: the vehicle parking processing method based on a security vehicle management terminal according to some embodiments of this disclosure can achieve real-time and accurate security monitoring of parked vehicles, and effectively prevent dangers to parked vehicles. Specifically, the reason for the inaccuracy of related security monitoring is that the parking vehicles rely on their own security devices to ensure their safety in real time, which requires real-time power supply to the security devices, resulting in a waste of resources. In addition, due to the limited security distance monitored by the parked vehicles, security loopholes still exist, leading to potential security risks for parked vehicles. Based on this, the vehicle parking processing method of some embodiments of this disclosure first uses a set of cameras installed in the target parking area to determine at least one vehicle information corresponding to at least one parked vehicle, so as to provide security protection for vehicles parked in the target parking area and prevent damage to the vehicles. Then, for each of the at least one vehicle information, the following sending steps are performed: First, determine whether the vehicle information is target vehicle information, so as to provide accurate security protection services for the registered vehicle information. Wherein, the target vehicle information is vehicle information that has been registered and authorized for parking security control. The second step involves obtaining the vehicle's parking location in response to the confirmed target vehicle information. This information is used to control the corresponding camera device for security monitoring. The third step involves identifying at least one camera device from the aforementioned camera device set that is adjacent to the vehicle's parking location. This information is then used to acquire a multi-directional video feed for real-time vehicle safety. The fourth step involves acquiring a multi-directional video feed of the vehicle information in real-time based on the at least one camera device information. This data serves as the foundation for accurately determining the hazard information of the target object. The fifth step involves accurately generating parking hazard information and the target object's intent information based on the multi-directional video feed. This allows for determining the degree of danger posed by the target object relative to the vehicle, thereby ensuring the safety of the vehicle. The target object is the object moving towards the vehicle's parking location. The sixth step involves, in response to determining that the value corresponding to the aforementioned parking hazard information falls within the first range, sending the parking hazard information and the aforementioned intent information to the vehicle's central control unit corresponding to the vehicle information, and sending a text message linking the aforementioned multi-directional video set and the aforementioned intent information to the user terminal of the registered user corresponding to the vehicle information. This avoids the problem of security vulnerabilities still existing due to the limited safe distance monitored for parked vehicles, thus preventing potential safety hazards associated with parked vehicles. Furthermore, it eliminates the need for parked vehicles to continuously power on the relevant safety devices, avoiding resource waste. Attached Figure Description

[0014] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0015] Figure 1 This is a flowchart of some embodiments of the vehicle parking management method based on a security vehicle management terminal according to the present disclosure;

[0016] Figure 2 This is a structural schematic diagram of some embodiments of a vehicle parking processing device based on a security vehicle management terminal according to the present disclosure;

[0017] Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0018] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure 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 disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0019] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0020] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0021] It should be noted that the terms "a" and "a plurality of" used in this disclosure 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 messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0023] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] refer to Figure 1The diagram illustrates a flow 100 of some embodiments of a vehicle parking processing method based on a security vehicle management terminal according to the present disclosure. The vehicle parking processing method includes the following steps:

[0025] Step 101: Using the set of cameras installed in the target parking area, determine at least one vehicle information corresponding to at least one parked vehicle.

[0026] In some embodiments, the entity executing the above-described vehicle parking processing method (e.g., a security vehicle management terminal) can utilize a set of cameras installed in the target parking area to determine at least one vehicle information corresponding to at least one parked vehicle. The target parking area can be a smart parking area. The cameras in the camera set can be infrared cameras installed in the parking lot. There is a one-to-one correspondence between the vehicles in the at least one vehicle and the vehicle information in the at least one vehicle information. The vehicle information can be a vehicle license plate. The smart parking security monitoring device can be a central control device corresponding to the target parking area. The security vehicle management terminal can be a terminal for performing vehicle security control in the smart parking lot.

[0027] Step 102: For each of the above at least one vehicle information, perform the following sending steps:

[0028] Step 1021: Determine whether the above vehicle information is the target vehicle information.

[0029] In some embodiments, the executing entity may determine whether the vehicle information is target vehicle information. The target vehicle information refers to vehicle information that has been registered and authorized for parking security control. That is, the vehicle corresponding to the target vehicle information is a vehicle that has been pre-registered and authorized for initial parking security control by the parking security monitoring device.

[0030] As an example, the aforementioned implementing entity can determine whether the vehicle information is the target vehicle information by querying the vehicle information registration form.

[0031] Step 1022: In response to determining that it is target vehicle information, obtain the vehicle parking location corresponding to the above vehicle information.

[0032] In some embodiments, in response to determining that it is target vehicle information, the executing entity can obtain the vehicle parking location corresponding to the vehicle information. The vehicle parking location can be the parking position of the vehicle corresponding to the vehicle information in the target parking area. The vehicle parking location can be a parking space number.

[0033] Step 1023: Determine information on at least one camera device corresponding to at least one camera device in the camera device cluster that is adjacent to the vehicle parking location.

[0034] In some embodiments, the executing entity may determine information about at least one camera device corresponding to at least one camera device in the camera device set that is adjacent to the vehicle parking location. Specifically, the device location in the target parking area corresponding to each camera device in the camera device set is predetermined. The cameras in the at least one camera device set may be cameras located in different orientations, and the distance between the device location corresponding to each camera device and the vehicle parking location is less than a predetermined distance. For example, the predetermined distance is 20 meters.

[0035] Step 1024: Based on the information from at least one of the above-mentioned camera devices, acquire a multi-angle video set of the above-mentioned vehicle information in real time.

[0036] In some embodiments, the executing entity can acquire a multi-angle video set of the vehicle information in real time based on the information from at least one camera device. There is a one-to-one correspondence between the camera device information in the at least one camera device information and the multi-angle videos in the multi-angle video set. The multi-angle videos can be videos captured specifically for the vehicle information.

[0037] As an example, the aforementioned executing entity can control at least one camera device to capture real-time video of vehicle information in order to generate a multi-view video set.

[0038] Step 1025: Based on the above multi-angle video set, generate parking hazard information and target object intent information.

[0039] In some embodiments, the executing entity can generate parking hazard information and target object intent information based on the multi-angle video set. The parking hazard information can characterize the degree of harm posed by the target object's behavior to the vehicle. The intent information can be the intended action of the target object. The target object is an object moving towards the vehicle's parking location.

[0040] In some optional implementations of certain embodiments, the aforementioned intent information includes at least one intent identification information. The intent identification information may be identified intent information. Each piece of intent identification information in the at least one intent identification information is different.

[0041] Optionally, generating parking hazard information and target object intent information based on the aforementioned multi-angle video set may include the following steps:

[0042] The first step is to perform the following processing steps for each multi-view video in the above multi-view video set:

[0043] The first sub-step involves preprocessing the aforementioned multi-directional video to generate a sequence of frame images.

[0044] As an example, firstly, the aforementioned execution entity can perform video decoding on multi-angle video to generate decoded video. Then, it performs frame extraction on the decoded video to generate a sequence of frame images.

[0045] The second sub-step involves performing the following generation steps for each frame in the above frame image sequence:

[0046] Sub-step 1 involves inputting the aforementioned frame image into a pre-trained target recognition model to generate object location information for the aforementioned target object and vehicle location information corresponding to the aforementioned vehicle information. The target recognition model can be an object detection model. In practice, the object detection model can be a YOLO model. The object location information can be the position information of the corresponding pixel of the target object within the frame image. The vehicle location information can be the position information of the corresponding pixel of the vehicle information within the frame image.

[0047] Sub-step 2: Based on the object location information and the vehicle location information, determine the outer bounding box. The outer bounding box can be a coordinate-based bounding box whose internal region includes both the object location information and the vehicle location information.

[0048] As an example, firstly, the executing entity can determine the size of the bounding box corresponding to the outer segmentation box. Then, using the object location information and vehicle location information as the center, a bounding box of the specified size is generated.

[0049] Sub-step 3: Based on the above-mentioned bounding box, crop the above-mentioned frame image to generate a cropped image.

[0050] As an example, the aforementioned execution entity can crop the aforementioned frame image based on the dividing lines of the outer segmentation box to generate a cropped image.

[0051] Sub-step 4 involves inputting the cropped image into a pre-trained image feature extraction model to generate image feature information. This image feature extraction model can be a neural network model for extracting image feature information. In practice, image feature information can characterize the semantic content of image features in the cropped image. The image feature extraction model can be a convolutional neural network model.

[0052] The third sub-step involves concatenating the obtained image feature information sequence in chronological order to obtain the first concatenated feature information.

[0053] The second step involves stitching together the first stitched feature information from the obtained first stitched feature information set according to a pre-set stitching method, to generate the second stitched feature information. For example, a multi-directional video set might correspond to a directional set including a first directional, a second directional, and a third directional. The stitching method could be to place the first stitched feature information corresponding to the first directional at the beginning, the first stitched feature information corresponding to the second directional in the middle, and the first stitched feature information corresponding to the third directional at the end.

[0054] The third step involves inputting the second concatenated feature information into a pre-trained intent recognition model to generate at least one intent recognition message. This intent recognition model can be a neural network model that generates the intent recognition message. For example, the intent recognition model could be a Transformer model.

[0055] The fourth step is to obtain the object registration information corresponding to the aforementioned target object. This object registration information can be the information registered by the target object in the intelligent parking safety monitoring device. The object registration information may include, but is not limited to, at least one of the following: a photo of the object, a comparison of the vehicle's license plate, and the object's contact information.

[0056] Fifth, in response to determining that the above-mentioned object registration information is empty, determine the parking hazard value range corresponding to each of the above-mentioned at least one intent identification information, thereby obtaining at least one parking hazard value range. The parking hazard value range can characterize the degree of harm of the intent identification information relative to the vehicle corresponding to the vehicle information.

[0057] As an example, the aforementioned executing entity can determine the parking hazard value range corresponding to each of the at least one intent identification information by using a relational table query.

[0058] Step 6: Determine the average of the at least one central value corresponding to the at least one parking hazard value interval as the parking hazard information. There is a one-to-one correspondence between the at least one parking hazard value interval and the at least one central value. The average value can be the value obtained by averaging the at least one central value.

[0059] Step 7: In response to determining that the above object registration information is not empty, determine the object movement route corresponding to the above target object. The object movement route can be the movement route of the target object within the target parking area.

[0060] As an example, the aforementioned execution entity can use a route extraction model to extract the motion route of the target object from a multi-view video dataset. The route extraction model can be a multi-layered, cascaded convolutional neural network model.

[0061] Step 8: Based on the above object movement route, select the intent recognition information with the highest probability of corresponding intent information from at least one intent recognition information, and use it as the target intent recognition information.

[0062] As an example, based on the object's movement route described above, key movement location information is determined. Then, from the at least one intent recognition information mentioned above, the intent recognition information with the highest location similarity to the key movement location information is selected as the target intent recognition information.

[0063] As another example, firstly, at least one predicted object movement route corresponding to the aforementioned at least one intent recognition information is determined. Then, from the at least one predicted object movement route, the predicted object movement route with the highest route similarity to the aforementioned object movement routes is selected as the target predicted object movement route. Finally, the intent recognition information corresponding to the target predicted object movement route is determined as the target intent recognition information.

[0064] Step 9: Determine the center value of the parking hazard value range corresponding to the above target intent identification information as the above parking hazard information.

[0065] In some optional implementations of certain embodiments, the second splicing feature information is input into a pre-trained intent recognition model to generate at least one intent recognition information, including the following steps:

[0066] The first step involves inputting the aforementioned second concatenated feature information into at least one cascaded convolutional neural network (CNN) model to generate an action recognition information sequence and at least one first initial intent recognition information. The at least one cascaded CNN model includes at least one cascaded CNN and at least one first action recognition information generation layer. The at least one cascaded CNN includes at least one target CNN. The positional information corresponding to the at least one target CNN is located in the middle network position of the at least one cascaded CNN. Each target CNN has a corresponding action recognition information generation layer. The action recognition information may include the corresponding recognition action of the target object and the action recognition probability. The action recognition probability can characterize the accuracy of the recognized action. There is a one-to-one correspondence between the target CNN and the action recognition information in the action recognition information sequence. The first action recognition information generation layer can be an action recognition information generation layer. In practice, the action recognition information generation layer can be a fully connected layer.

[0067] The second step is to combine the action recognition information that is the same and adjacent in the above action recognition information sequence to generate at least one action recognition information subsequence.

[0068] Third, for at least one action recognition information subsequence, perform the following information generation steps:

[0069] The first sub-step involves determining the target convolutional neural network corresponding to each action recognition information in the above action recognition information sub-sequence, thereby obtaining the target convolutional neural network sequence.

[0070] The second sub-step is to determine the convolutional network output information sequence corresponding to the target convolutional neural network sequence.

[0071] The third sub-step involves concatenating the output information of each convolutional network in the above sequence of convolutional network output information to generate concatenated output information.

[0072] The fourth sub-step involves inputting the spliced ​​output information into the second action recognition information generation layer to generate joint action recognition information.

[0073] The fifth sub-step involves determining that the action recognition information included in the joint action recognition information is the same as each action recognition information in the action recognition information sub-sequence, and that the included action recognition probability is greater than any action recognition probability in the corresponding action recognition probability sub-sequence of the action recognition information sub-sequence, and then determining the action recognition information in the aforementioned action recognition information sub-sequence as the execution action information corresponding to the target object.

[0074] The fourth step is to generate at least one second initial intent information based on the obtained at least one execution action information by querying the intent association table.

[0075] The fifth step is to concatenate the at least one first initial intent information and the at least one second initial intent information to generate at least one intent recognition information.

[0076] The content described above in "some optional implementations in some embodiments" serves as an inventive point of this disclosure, solving the technical problem mentioned in the background art: "The accuracy of intent recognition affects vehicle safety, and the accuracy of intent recognition is an aspect that needs to be accurately guaranteed." Based on this, this disclosure, through an intent recognition model including at least one cascaded convolutional neural network model, uses both sequential actions and overall feature information to accurately generate at least one intent recognition information for a target object.

[0077] Step 1026: In response to determining that the value corresponding to the above-mentioned parking hazard information is in the first interval, the above-mentioned parking hazard information and the above-mentioned intent information are sent to the vehicle central control device corresponding to the above-mentioned vehicle information, and a link SMS for the above-mentioned multi-directional video set and the above-mentioned intent information is sent to the user terminal of the registered object corresponding to the above-mentioned vehicle information.

[0078] In some embodiments, in response to determining that the value corresponding to the parking hazard information falls within a first range, the executing entity may send the parking hazard information and the intent information to the in-vehicle central control device corresponding to the vehicle information, and send a link SMS message for the multi-directional video set and the intent information to the user terminal of the registered object corresponding to the vehicle information. For example, the first range may be [70-100]. The in-vehicle central control device may be an intelligent central control device for the vehicle lock device corresponding to the vehicle information. The registered object may be the object registered to the vehicle corresponding to the vehicle information. The link SMS message includes: a compressed link for the multi-directional video set and the intent information, and a hazard warning message. The user terminal may be the mobile terminal used by the registered object.

[0079] In some optional implementations of certain embodiments, after step 1026, the steps further include:

[0080] The first step involves, in response to determining that the value corresponding to the aforementioned parking hazard information falls within the second interval and receiving vehicle processing information sent to the aforementioned registered object, comparing the vehicle processing information with the aforementioned intent information to obtain comparison information. The value in the aforementioned first interval is greater than the value in the aforementioned second interval. For example, the second interval could be [50-70]. The comparison information includes: information indicating consistent representation and information indicating inconsistent representation.

[0081] The second step, in response to the confirmation that the aforementioned comparison information represents the same information, is to modify the numerical value corresponding to the aforementioned parking hazard information to the target value, and to package and store the aforementioned multi-directional video set, vehicle processing information, and intent information. In practice, the target value can be the value 20. The target value corresponds to a value smaller than the value in the second interval.

[0082] Thirdly, in response to determining that the value corresponding to the aforementioned parking hazard information falls within the aforementioned second range and no vehicle processing information has been received, the aforementioned parking hazard information is sent to the aforementioned in-vehicle central control device, so that the aforementioned in-vehicle central control device can activate the in-vehicle safety monitoring system. The in-vehicle safety monitoring system may be a monitoring system pre-installed on the vehicle corresponding to the vehicle information to ensure vehicle safety.

[0083] Optionally, in response to determining that the above comparison information characterization information is inconsistent, the above intent information and the above vehicle processing information are sent to the user terminal for authenticity verification.

[0084] In some optional implementations of certain embodiments, the process of comparing the vehicle processing information and the intent information to obtain comparison information may include the following steps:

[0085] The first step is to obtain a pre-built intent information tree. This intent information tree can be a pre-built tree where each node represents a specific intent. The closer a node is to the root node, the broader the range of its corresponding intent information. For example, node A might correspond to the intent "walk towards the target vehicle." The next level node under node A might correspond to the intent "walk towards the target vehicle and drive the target vehicle." This intent information tree can be a pre-constructed tree.

[0086] The second step is to determine the intent information corresponding to the above vehicle processing information, which will be used as the target intent information.

[0087] As an example, firstly, the aforementioned executing entity can perform word segmentation on the vehicle processing information to obtain a word set. Then, based on the word set, it can use an intent recognition model to generate target intent information.

[0088] The third step is to determine the intent information subtree rooted at the intent information in the intent information tree.

[0089] The fourth step is to determine whether there exists any intent information in the aforementioned intent information subtree with an intent similarity greater than the target intent information. The intent similarity can be the cosine similarity between intent vectors. The target similarity value can be 0.85.

[0090] The fifth step is to generate matching information that is consistent with the identified information in response to the confirmation of existence.

[0091] In some optional implementations of certain embodiments, the above-mentioned vehicle central control device performs hazard identification through the following steps:

[0092] The first step involves, in response to receiving the aforementioned parking hazard information and intent information from the intelligent parking safety monitoring device, performing a power-on process for at least one vehicle-mounted camera and a LiDAR device. The power-on process may be a power supply process. The vehicle-mounted camera may be a camera installed on the vehicle corresponding to the vehicle information. The LiDAR device is used to acquire a point cloud dataset of the area surrounding the vehicle corresponding to the vehicle information.

[0093] The second step involves using at least one vehicle-mounted camera device to acquire at least one real-time external video feed, and using the aforementioned LiDAR device to acquire an external point cloud dataset. There is a one-to-one correspondence between the real-time external video feed and the vehicle-mounted camera device in the at least one vehicle-mounted camera setup.

[0094] The third step is to generate object intent information for the target object based on at least one real-time external video and the external point cloud dataset.

[0095] The fourth step is to determine the hazard level corresponding to the aforementioned object intent information. The hazard level characterizes the degree of harm that the object intent information may cause to the vehicle corresponding to the vehicle information. The hazard level can be one of the following: Level 1 and Level 2. The hazard level corresponding to Level 1 is less than the hazard level corresponding to Level 2.

[0096] Fifth, in response to determining the aforementioned danger level as Level 1, continue controlling at least one vehicle-mounted camera and the aforementioned lidar device to perform video capture and point cloud scanning of the target object until the distance between the target object and the vehicle corresponding to the aforementioned vehicle information is greater than the target distance. In practice, the target distance can be a pre-set distance.

[0097] Step 6: In response to determining that the above-mentioned danger level is Level 2, control the voice playback device to play a predetermined script, and continue to control the above-mentioned at least one vehicle-mounted camera device and the above-mentioned lidar device to perform video shooting and point cloud scanning on the above-mentioned target object until the distance between the above-mentioned target object and the vehicle corresponding to the above-mentioned vehicle information is greater than the target distance.

[0098] Step 7: In response to determining that the aforementioned danger level is Level 2 and the filming duration for the aforementioned target object exceeds the target duration, a warning link SMS is sent to the user terminal corresponding to the registered object, and the filmed video set for the aforementioned target object is uploaded in real time to the target application on the user terminal corresponding to the registered object. The link in the warning link SMS is a target application redirection link. The target application redirection link can be a link that redirects to the target application, allowing the registered object to view the filmed video set.

[0099] Step 8: In response to determining the above hazard level as Level 3, perform the following procedures:

[0100] Operation 1: Control the voice playback device to play the pre-reserved script.

[0101] Operation 2: Execute the alarm processing for the above vehicle information.

[0102] Operation 3: Send the generated object danger information to the aforementioned intelligent parking safety monitoring device.

[0103] Step 4: Send a telephone alert to the terminal corresponding to the registered object and upload the video set of the target object to the target application in real time.

[0104] Optionally, generating object intent information for the target object based on at least one real-time exterior video and the exterior point cloud dataset may include the following steps:

[0105] The first step involves generating at least one first candidate object intent information based on at least one real-time external video feed, using an image-based intent recognition model. This image-based intent recognition model can be generated based on input information in image form. In practice, the image-based intent recognition model can be a multi-layered, cascaded encoding and decoding neural network model.

[0106] As an example, firstly, the aforementioned executing entity can generate overall exterior feature information for at least one real-time exterior video. Then, the overall exterior feature information is input into an image-based intent recognition model to generate intent information for at least one first candidate object.

[0107] The second step involves generating at least one second candidate object's intent information based on the aforementioned vehicle exterior point cloud dataset and a point cloud-based intent recognition model. This point cloud-based intent recognition model can be generated based on point cloud-based input information. In practice, it can be a multi-head attention mechanism combined with a convolutional neural network model.

[0108] As an example, firstly, the aforementioned execution entity can generate overall exterior point cloud feature information for the vehicle exterior point cloud dataset. Then, the overall exterior point cloud feature information is input into an intent recognition model based on point cloud format to generate at least one second candidate object intent information.

[0109] The third step is to fuse the intent information of at least one first candidate object and the intent information of at least one second candidate object to generate a set of candidate object intent information.

[0110] The fourth step is to deduplicate the above candidate object intent information set to obtain the deduplicated object intent information set.

[0111] The fifth step involves defining the deduplicated object intent information set and its corresponding intent information weight set as the object intent information. There is a one-to-one correspondence between the intent information weights in the intent information weight set and the object intent information in the deduplicated object intent information set. The intent information weights characterize the importance of the corresponding intent information. These weights can be generated by normalizing based on the repetition frequency of the object intent information in the candidate object intent information set.

[0112] In some optional implementations of certain embodiments, after step 1027, the steps further include:

[0113] In response to receiving the aforementioned danger information from the vehicle central control device, the executing entity can use the information of at least one camera device corresponding to at least one camera device to zoom in and take pictures of the target object and the vehicle corresponding to the vehicle information, thereby obtaining a magnified video of the target object, and control the audio playback device corresponding to the parking position of the vehicle to play a warning message for the target object and the vehicle information.

[0114] The various embodiments of this disclosure have the following beneficial effects: the vehicle parking handling method of some embodiments of this disclosure can achieve real-time and accurate safety monitoring of parked vehicles, and effectively prevent dangers to parked vehicles. Specifically, the reason for the inaccuracy of related safety monitoring is that the safety protection devices of parked vehicles need to be powered on in real time, which leads to a waste of resources. In addition, due to the limited safety distance of the parked vehicles, there are still security loopholes, resulting in potential safety hazards for parked vehicles. Based on this, the vehicle parking handling method of some embodiments of this disclosure first uses a set of camera devices installed in the target parking area to determine at least one vehicle information corresponding to at least one parked vehicle, so as to provide security protection for vehicles parked in the target parking area and prevent damage to vehicles. Then, for each of the at least one vehicle information, the following sending steps are performed: First, determine whether the vehicle information is target vehicle information, so as to provide accurate security protection services for the registered vehicle information. Wherein, the target vehicle information is vehicle information that has been registered and authorized for parking security control. The second step involves obtaining the vehicle's parking location in response to the confirmed target vehicle information. This information is used to control the corresponding camera device for security monitoring. The third step involves identifying at least one camera device from the aforementioned camera device set that is adjacent to the vehicle's parking location. This information is then used to acquire a multi-directional video feed for real-time vehicle safety. The fourth step involves acquiring a multi-directional video feed of the vehicle information in real-time based on the at least one camera device information. This data serves as the foundation for accurately determining the hazard information of the target object. The fifth step involves accurately generating parking hazard information and the target object's intent information based on the multi-directional video feed. This allows for determining the degree of danger posed by the target object relative to the vehicle, thereby ensuring the safety of the vehicle. The target object is the object moving towards the vehicle's parking location. The sixth step involves, in response to determining that the value corresponding to the aforementioned parking hazard information falls within the first range, sending the parking hazard information and the aforementioned intent information to the vehicle's central control unit corresponding to the vehicle information, and sending a text message linking the aforementioned multi-directional video set and the aforementioned intent information to the user terminal of the registered user corresponding to the vehicle information. This avoids the problem of security vulnerabilities still existing due to the limited safe distance monitored for parked vehicles, thus preventing potential safety hazards associated with parked vehicles. Furthermore, it eliminates the need for parked vehicles to continuously power on the relevant safety devices, avoiding resource waste.

[0115] Further reference Figure 2As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a vehicle parking processing device, which are similar to... Figure 1 Corresponding to the method embodiments shown, this vehicle parking processing device can be specifically applied to various electronic devices.

[0116] like Figure 2 As shown, a vehicle parking processing device 200 includes a determining unit 201 and an execution unit 202. The determining unit 201 is configured to use a set of cameras installed in a target parking area to determine at least one vehicle information corresponding to at least one parked vehicle. The execution unit 202 is configured to perform the following sending steps for each of the at least one vehicle information: determining whether the vehicle information is target vehicle information, wherein the target vehicle information is vehicle information that has been registered and authorized for parking security control; in response to determining that it is target vehicle information, obtaining the vehicle parking location corresponding to the vehicle information; and determining at least one vehicle from the set of cameras adjacent to the vehicle parking location. Information on at least one camera device corresponding to the camera device; real-time acquisition of a multi-directional video set of the vehicle information based on the at least one camera device information; generation of parking hazard information and target object intent information based on the multi-directional video set, wherein the target object is an object moving toward the parking location of the vehicle; in response to determining that the value corresponding to the parking hazard information is in a first interval, sending the parking hazard information and the intent information to the in-vehicle central control device corresponding to the vehicle information, and sending a link SMS message for the multi-directional video set and the intent information to the user terminal of the registered object corresponding to the vehicle information.

[0117] It is understandable that the units described in the vehicle parking processing device 200 are similar to those in the reference device. Figure 1 The steps in the described method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method also apply to the vehicle parking processing device 200 and the units contained therein, and will not be repeated here.

[0118] The following is for reference. Figure 3 It shows a schematic diagram of the structure of an electronic device (e.g., an electronic device) 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0119] like Figure 3As shown, the electronic device 300 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0120] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0121] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0122] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0123] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0124] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: use a set of cameras installed at the target parking area to determine at least one vehicle information corresponding to at least one parked vehicle; for each of the at least one vehicle information, perform the following transmission steps: determine whether the vehicle information is target vehicle information, wherein the target vehicle information is vehicle information that has been registered and authorized for parking security control; in response to determining that it is target vehicle information, obtain the vehicle parking location corresponding to the vehicle information; determine the adjacent, upper... The system includes at least one camera device information corresponding to at least one camera device in the camera device set; based on the at least one camera device information, it acquires a multi-directional video set of the vehicle information in real time; based on the multi-directional video set, it generates parking hazard information and target object intent information, wherein the target object is an object moving towards the vehicle parking location; in response to determining that the value corresponding to the parking hazard information is in a first interval, it sends the parking hazard information and the intent information to the vehicle central control device corresponding to the vehicle information, and sends a link SMS message for the multi-directional video set and the intent information to the user terminal of the registered user corresponding to the vehicle information.

[0125] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0126] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0127] The units described in some embodiments of this disclosure can be implemented in software or in hardware. The described units can also be housed in a processor; for example, a processor may be described as including a determining unit and an executing unit. The names of these units do not necessarily limit the unit itself; for example, a determining unit may also be described as "a unit that determines at least one vehicle information corresponding to at least one parked vehicle using a set of cameras installed in a target parking area."

[0128] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0129] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A vehicle parking management method based on a security vehicle management terminal, comprising: Using a set of cameras installed in the target parking area, determine at least one vehicle information corresponding to at least one parked vehicle; For each of the at least one vehicle information, perform the following sending steps: Determine whether the vehicle information is the target vehicle information, wherein the target vehicle information is the vehicle information that has been registered and authorized for parking safety control; In response to determining that it is target vehicle information, the vehicle parking location corresponding to the vehicle information is obtained; Determine at least one camera device information corresponding to at least one camera device in the camera device set that is adjacent to the vehicle parking location, wherein each camera device in the at least one camera device is a camera device in a different orientation, and the positional distance between the device position corresponding to each camera device and the vehicle parking location is less than a preset distance; Based on the information from the at least one camera device, a multi-angle video set targeting the vehicle information is acquired in real time. Based on the multi-directional video set, generate second stitching feature information; Based on the second concatenation feature information, at least one intent recognition information is generated, including: The second spliced ​​feature information is input into at least one cascaded convolutional neural network model to generate an action recognition information sequence and at least one first initial intent recognition information; The action recognition information that is the same and adjacent in the action recognition information sequence is combined to generate at least one action recognition information subsequence; Determine the target convolutional neural network corresponding to each action recognition information in the action recognition information sub-sequence to obtain the target convolutional neural network sequence; Determine the convolutional network output information sequence corresponding to the target convolutional neural network sequence; The output information of each convolutional network in the convolutional network output information sequence is concatenated to generate concatenated output information; The spliced ​​output information is input into the second action recognition information generation layer to generate joint action recognition information; In response to determining that the action recognition information included in the joint action recognition information is the same as each action recognition information in the action recognition information subsequence, and the included action recognition probability is greater than any action recognition probability in the corresponding action recognition probability subsequence of the action recognition information subsequence, the action recognition information in the action recognition information subsequence is determined as the execution action information corresponding to the target object; Based on the obtained at least one execution action information, at least one second initial intent information is generated by querying the intent association table; The at least one first initial intent information and the at least one second initial intent information are concatenated to generate intent recognition information with one less element. Based on at least one intent recognition information, parking hazard information and target object intent information are generated, wherein the target object is an object moving toward the vehicle parking location; In response to determining that the value corresponding to the parking hazard information is in the first range, the parking hazard information and the intent information are sent to the in-vehicle central control device corresponding to the vehicle information, and a link SMS message for the multi-directional video set and the intent information is sent to the user terminal of the registered object corresponding to the vehicle information.

2. The method according to claim 1, wherein, After responding to determining that the value corresponding to the parking hazard information is within a first interval, sending the parking hazard information and the intent information to the in-vehicle central control device corresponding to the vehicle information, and sending a linked SMS message for the multi-directional video set and the intent information to the user terminal of the registered object corresponding to the vehicle information, the method further includes: In response to determining that the value corresponding to the parking danger information is in the second interval and receiving vehicle processing information sent to the registered object, the vehicle processing information and the intention information are compared to obtain comparison information, wherein the value in the first interval is greater than the value in the second interval; In response to determining that the comparison information characterization information is consistent, the value corresponding to the parking hazard information is modified to the target value, and the multi-directional video set, the vehicle processing information and the intent information are packaged and stored. In response to determining that the value corresponding to the parking hazard information is in the second range and no vehicle processing information has been received, the parking hazard information is sent to the vehicle central control device so that the vehicle central control device can activate the vehicle safety monitoring system.

3. The method according to claim 2, wherein, The step of comparing the vehicle processing information and the intent information to obtain comparison information includes: Obtain a pre-built intent information tree; Determine the intent information corresponding to the vehicle processing information as the target intent information; Determine the intent information subtree rooted at the intent information in the intent information tree; Determine whether there exists any intent information in the intent information subtree whose intent similarity to the target intent information is greater than the target similarity value; In response to the confirmation of existence, comparison information with consistent representational information is generated.

4. The method according to claim 1, wherein, The intent information includes: at least one intent identification information; and The step of generating parking hazard information and target object intent information based on the multi-angle video set includes: For each multi-view video in the multi-view video set, perform the following processing steps: The multi-directional video is preprocessed to generate a sequence of frame images; For each frame image in the frame image sequence, perform the following generation steps: The frame image is input into a pre-trained target recognition model to generate object location information for the target object and vehicle location information corresponding to the vehicle information; The outer bounding box is determined based on the object location information and the vehicle location information; Based on the bounding box, the frame image is cropped to generate a cropped image; The cropped image is input into a pre-trained image feature extraction model to generate image feature information; The obtained image feature information sequence is concatenated in chronological order to obtain the first concatenated feature information; Based on a pre-set directional feature information splicing method, each first splicing feature information in the obtained first splicing feature information set is spliced ​​together to generate second splicing feature information; The second splicing feature information is input into a pre-trained intent recognition model to generate at least one intent recognition information; Obtain the object registration information corresponding to the target object; In response to determining that the object registration information is empty, the parking hazard value range corresponding to each of the at least one intent identification information is determined to obtain at least one parking hazard value range; The average value of at least one interval center value corresponding to the at least one parking hazard value interval is determined as the parking hazard information; In response to determining that the object registration information is not empty, the object movement route corresponding to the target object is determined; Based on the object's movement path, the intent recognition information with the highest probability of corresponding intent information is selected from the at least one intent recognition information and used as the target intent recognition information; The center value of the parking hazard value range corresponding to the target intent identification information is determined as the parking hazard information.

5. The method according to claim 1, wherein, The vehicle-mounted central control device identifies hazards through the following steps: In response to receiving the parking danger information and the intent information sent by the intelligent parking safety monitoring device, power-on processing is performed for at least one vehicle-mounted camera device and a lidar device; Using the at least one vehicle-mounted camera device, acquire at least one real-time video of the vehicle exterior; and using the lidar device, acquire a point cloud dataset of the vehicle exterior. Based on the at least one real-time external video and the external point cloud dataset, generate object intent information for the target object; Determine the danger level corresponding to the object's intent information; In response to determining the danger level as Level 1, the at least one vehicle-mounted camera device and the lidar device continue to be controlled to perform video capture and point cloud scanning on the target object until the distance between the target object and the vehicle corresponding to the vehicle information is greater than the target distance. In response to determining that the danger level is Level 2, the voice playback device is controlled to play a predetermined script, and the at least one vehicle-mounted camera device and the lidar device are continued to be controlled to perform video capture and point cloud scanning on the target object until the distance between the target object and the vehicle corresponding to the vehicle information is greater than the target distance. In response to determining that the danger level is the second level and the shooting duration for the target object is longer than the target duration, a warning link SMS is sent to the user terminal corresponding to the registered object, and the video set of the target object is uploaded in real time to the target application in the user terminal corresponding to the registered object, wherein the link in the warning link SMS is a target application jump link; In response to determining that the hazard level is Level 3, the following processing operations are performed: Control the voice playback device to play the pre-set script; Perform alarm processing for the vehicle information; Send the generated object danger information to the intelligent parking safety monitoring device; A telephone alert is sent to the terminal corresponding to the registered object, and a set of videos taken of the target object is uploaded to the target application in real time.

6. The method according to claim 5, wherein, After responding to determining that the value corresponding to the parking hazard information is within a first interval, sending the parking hazard information and the intent information to the in-vehicle central control device corresponding to the vehicle information, and sending a linked SMS message for the multi-directional video set and the intent information to the user terminal of the registered object corresponding to the vehicle information, the method further includes: In response to receiving the object danger information sent by the vehicle central control device, the system uses the at least one camera device information corresponding to at least one camera device to zoom in and capture images of the target object and the vehicle information corresponding to the vehicle to obtain a zoomed-in video of the object, and controls the audio playback device corresponding to the vehicle parking position to play a warning message for the target object and the vehicle information.

7. The method according to claim 5, wherein, The step of generating object intent information for the target object based on the at least one real-time exterior video and the exterior point cloud dataset includes: Based on the at least one real-time external video, at least one first candidate object intent information is generated using an image-based intent recognition model; Based on the aforementioned vehicle exterior point cloud dataset, at least one second candidate object intent information is generated using a point cloud-based intent recognition model. The intent information of at least one first candidate object and the intent information of at least one second candidate object are fused together to generate a set of candidate object intent information; The candidate object intent information set is deduplicated to obtain the deduplicated object intent information set; The deduplicated set of object intent information and the corresponding set of intent information weights are determined as object intent information.

8. A vehicle parking management device based on a security vehicle management terminal, comprising: The determining unit is configured to use a set of cameras installed in the target parking area to determine at least one vehicle information corresponding to at least one parked vehicle; An execution unit is configured to perform the following sending steps for each of the at least one vehicle information: determining whether the vehicle information is target vehicle information, wherein the target vehicle information is vehicle information that has been registered and authorized for parking safety control; in response to determining that it is target vehicle information, obtaining the vehicle parking location corresponding to the vehicle information; determining at least one camera device information corresponding to at least one camera device in the camera device set that is adjacent to the vehicle parking location, wherein each camera device in the at least one camera device is a camera device in a different orientation, and the positional distance between the device location corresponding to each camera device and the vehicle parking location is less than a preset distance; according to The at least one camera device acquires a multi-directional video set of the vehicle information in real time; generates second stitching feature information based on the multi-directional video set; and generates at least one intent recognition information based on the second stitching feature information, including: inputting the second stitching feature information into at least one cascaded convolutional neural network model to generate an action recognition information sequence and at least one first initial intent recognition information; combining action recognition information that is identical and adjacent in the action recognition information sequence to generate at least one action recognition information sub-sequence; determining the target convolutional neural network corresponding to each action recognition information in the action recognition information sub-sequence to obtain a target convolutional neural network sequence; and determining... The target convolutional neural network sequence corresponds to the convolutional network output information sequence; the output information of each convolutional network in the convolutional network output information sequence is concatenated to generate concatenated output information; the concatenated output information is input to the second action recognition information generation layer to generate joint action recognition information; in response to determining that the action recognition information included in the joint action recognition information is the same as each action recognition information in the action recognition information subsequence, and the included action recognition probability is greater than any action recognition probability in the corresponding action recognition probability subsequence of the action recognition information subsequence, the action recognition information in the action recognition information subsequence is determined as the execution action information corresponding to the target object; based on the obtained at least one execution action information The system generates at least one second initial intent information by querying an intent association table; it then concatenates the at least one first initial intent information and the at least one second initial intent information to generate at least one intent recognition information; based on the at least one intent recognition information, it generates parking hazard information and intent information for a target object, wherein the target object is an object moving toward the vehicle's parking location; in response to determining that the value corresponding to the parking hazard information is within a first interval, it sends the parking hazard information and the intent information to the in-vehicle central control device corresponding to the vehicle information, and sends a link SMS message for the multi-directional video set and the intent information to the user terminal of the registered object corresponding to the vehicle information.

9. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images