Detection method and device based on stereo garage

Abstract

The invention relates to the technical field of intelligent parking garage management, in particular to a detection method and device based on a stereo garage, and the method comprises the steps: receiving a starting instruction of a target vehicle in a preset stereo garage area; performing multidirectional image acquisition on the target vehicle according to a preset image acquisition frequency to obtain an image set of the target vehicle in multiple directions; performing image preprocessing on the image set of the target vehicle to obtain a preprocessed target image set of the target vehicle; performing image detection on the target image set to obtain a target object of each target image in the target image set; tracking the target object of each target image, and determining whether the target object of each target image is in an abnormal state; and when the target object of each target image is not in the abnormal state, the target vehicle in the preset stereo garage area is parked through the starting instruction. According to the method, the accuracy and the flexibility based on stereo garage detection are improved.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent parking garage management, and in particular to a detection method and device based on a three-dimensional parking garage. Background Art

[0002] In today's society, with the acceleration of urbanization, urban space resources have become increasingly scarce. At the same time, the surge in vehicle ownership has made parking a major problem in urban management. In order to effectively alleviate this situation, three-dimensional parking garages have been widely used and promoted as an innovative parking solution.

[0003] Despite this, the existing three-dimensional parking garage management system still faces a series of technical bottlenecks in the detection of personnel and objects. First, the limitation of the monitoring angle of view is a prominent problem. Due to the complex structure of the three-dimensional parking garage, the existing monitoring system is difficult to achieve full coverage of every corner of the garage, resulting in some areas becoming monitoring blind spots, and safety hazards cannot be effectively prevented. Secondly, the quality of image acquisition is limited by many factors. Environmental factors such as light intensity and shooting angle have a direct impact on the image acquisition effect, making the captured target features unclear, increasing the difficulty of target detection, and thus affecting the accurate judgment and timely response capabilities of abnormal situations. Thirdly, traditional detection methods perform poorly in terms of accuracy and real-time performance, and thus it is difficult to meet the needs of real-time monitoring. Potential safety hazards and abnormal situations cannot be discovered and handled in a timely manner, thus affecting the overall safety performance and operational efficiency of the three-dimensional parking garage.

[0004] In summary, the existing monitoring methods of three-dimensional parking garages still have a lot of room for improvement in terms of ensuring garage safety and improving management level. How to overcome these technical difficulties and achieve more efficient and accurate monitoring and management has become an urgent problem to be solved in the current development of three-dimensional parking garages. Summary of the invention

[0005] One purpose of an embodiment of the present invention is to provide a detection method and device based on a three-dimensional parking garage, which is used to solve the technical problems existing in the three-dimensional parking garage, such as incomplete monitoring coverage, low image acquisition quality, and insufficient accuracy and timeliness of the detection method.

[0006] In a first aspect, an embodiment of the present invention provides a detection method based on a stereoscopic parking garage, which is applied to a detection system of a stereoscopic parking garage, and the method includes:

[0007] Receiving a start instruction for a target vehicle in a preset three-dimensional parking garage area, the start instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area;

[0008] According to a preset image acquisition frequency, multi-directional image acquisition is performed on the target vehicle to obtain an image set of the target vehicle in multiple directions;

[0009] Performing image preprocessing on the image set of the target vehicle to obtain a preprocessed target image set of the target vehicle;

[0010] Performing image detection on the target image set to obtain a target object of each target image in the target image set;

[0011] Tracking the target object of each target image, and determining whether the target object of each target image is in an abnormal state;

[0012] When the target object of each target image is not in an abnormal state, the start instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area.

[0013] In a second aspect, an embodiment of the present invention provides a detection device based on a stereo garage, which is applied to a detection system of a stereo garage, and the device includes:

[0014] A receiving unit, configured to receive a confirmation instruction of a target vehicle located in a preset three-dimensional parking garage area, wherein the confirmation instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area;

[0015] An acquisition unit, used to acquire images of the target vehicle according to a preset image acquisition frequency to obtain an image set of the target vehicle;

[0016] A processing unit, used for performing image preprocessing on the image set of the target vehicle to obtain a preprocessed target image set of the target vehicle;

[0017] A detection unit, configured to perform image detection on the target image set to obtain a target object of each target image in the target image set;

[0018] a determination unit, configured to track a target object in each target image and determine whether the target object in each target image is in an abnormal state;

[0019] The determining unit is further configured to, when the target object of each target image is not in an abnormal state, perform a parking operation on the target vehicle in the preset three-dimensional parking garage area through the start instruction.

[0020] In a third aspect, an embodiment of the present invention provides a computer device, including:

[0021] at least one processor; and,

[0022] a memory communicatively connected to the at least one processor; wherein,

[0023] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to the first aspect.

[0024] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the method as described in the first aspect.

[0025] In the scheme implemented by the above-mentioned detection method, device, equipment and storage medium based on the three-dimensional parking garage, the method reduces the link of human judgment through automated image acquisition and processing, reduces errors caused by operator negligence, and improves the reliability of the system. Through continuous image acquisition and detection, the system can monitor the situation in the garage in real time, and respond quickly to any abnormal situation and take timely measures; further, through real-time monitoring and tracking of the image of the target vehicle, the system can promptly detect abnormal conditions, such as vehicle position deviation, abnormal stop, etc., thereby avoiding potential collisions and accidents and ensuring the safety of vehicles and personnel in the garage; and the system can perform parking operations through start instructions after confirming that the vehicle is safe, reducing human intervention and improving the automation level and parking efficiency of the three-dimensional parking garage; and the system can be adjusted according to different garage structures and vehicle types, with good adaptability and flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0027] Figure 1 It is a schematic diagram of a detection system of a stereo garage in one embodiment of the present invention;

[0028] Figure 2 It is a flow chart of a detection method based on a stereoscopic car park in one embodiment of the present invention;

[0029] Figure 3 It is a structural schematic diagram of a detection device based on a three-dimensional parking garage in one embodiment of the present invention;

[0030] Figure 4 It is a schematic diagram of the structure of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0032] It should be noted that, if there is no conflict, the various features in the embodiments of the present invention can be combined with each other, all within the scope of protection of the present invention. In addition, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in a different order from the module division in the device or the flow chart. Furthermore, the words "first", "second", "third", etc. used in the present invention do not limit the data and execution order, but only distinguish the same items or similar items with basically the same functions and effects.

[0033] The technical solution of the present application can be applied to various three-dimensional parking scenarios.

[0034] In the detection method based on the stereo garage, the detection system applied to the stereo garage is shown in Figure 1 , Figure 1 A schematic diagram of a detection system for a three-dimensional parking garage provided in an embodiment of the present application, wherein the detection system 10 for the three-dimensional parking garage includes a camera module 20 , a transmission module 30 , a processing module 40 , and a notification module 50 .

[0035] Among them, the camera module 20 adopts a camera with high frame rate and high resolution, which can obtain clear and detailed images in a short time; and the camera module 20 is equipped with an automatic adjustment function, which can adjust the focal length, exposure time and other parameters in real time according to the changes in ambient light to ensure the stability of image quality.

[0036] Among them, the camera module 20 can be arranged at the entrance and exit of the multi-story parking garage to capture the entry and exit of vehicles and personnel; and cameras are arranged at intervals on the top and both sides of the passages on each floor to ensure all-round monitoring of people and vehicles in the passages.

[0037] The transmission module 30 uses a high-speed wired network (such as Ethernet) or advanced wireless network technology (such as 5G) to ensure that the image data can be quickly and without delay transmitted to the processing module 40. At the same time, data encryption and error correction technology are used in the transmission module 30 to ensure the security and integrity of the data during the transmission process.

[0038] The processing module 40 includes an image processing and analysis submodule and a database submodule. The image processing and analysis submodule is used for image preprocessing, target detection, target classification, and behavior analysis. The database submodule is used to store a large amount of original images, processed images, detection results, and related vehicle and personnel information, and adopts a distributed storage architecture to improve data storage efficiency and access speed, and support long-term storage and rapid retrieval of data.

[0039] Specifically, image preprocessing is to use a variety of algorithms to perform denoising, enhancement, correction and other operations on the original image, remove interference factors in the image, and highlight target features; target detection is a target detection algorithm based on deep learning, which can quickly and accurately identify people, vehicles and objects in the image, and determine their positions and contours; target classification is to perform fine classification of detected targets, such as distinguishing different types of vehicles, gender and age of people, etc.; behavior analysis is to determine whether the behavior of people is normal through continuous tracking and motion trajectory analysis of the target, such as whether they are walking in the specified area, whether there is any abnormal stay, etc.

[0040] The notification module 50 is used to send out alarms or reminders in abnormal situations. When the system detects an abnormal situation, such as people being trapped or vehicles exceeding the limit, the sound and light alarm device is immediately triggered, the garage stops running, and detailed notification information is sent to computers and other terminals; it supports multiple alarm methods, such as voice, application push, etc., to ensure that managers can receive and respond in time.

[0041] In this solution, the processing module 40 in the detection system 10 of the three-dimensional parking garage receives a start instruction of a target vehicle located in a preset three-dimensional parking garage area, and the start instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area; the start instruction at this time can be transmitted to the processing module 40 through the transmission module 30; the detection system 10 of the three-dimensional parking garage controls the camera module 20 to perform multi-directional image acquisition of the target vehicle according to a preset image acquisition frequency, and obtains an image set of the target vehicle in multiple directions; the camera module 20 transmits the acquired data to the processing module 40 through the transmission module 30, and the processing module 40 processes the image set of the target vehicle The processing module 40 performs image detection on the target image set to obtain the target object of each target image in the target image set; the processing module 40 tracks the target object of each target image to determine whether the target object of each target image is in an abnormal state; when the target object of each target image is not in an abnormal state, the processing module 40 performs a parking operation on the target vehicle in the preset stereoscopic parking garage area through the start instruction; if it is in an abnormal state, the processing module 40 sends a notification to the notification module 50, and the notification module 50 can alarm or remind according to the instruction.

[0042] In summary, through the collaborative work of these four core modules, the detection system 10 of the three-dimensional parking garage provides a comprehensive, efficient and reliable monitoring and security solution for garage management, which not only greatly improves the safety level of the garage, but also optimizes the management efficiency and user experience.

[0043] In view of this, the present application proposes a detection method based on a stereoscopic car park to solve the above problems, which is described in detail below.

[0044] See also Figure 2 , Figure 2 A flow chart of a detection method based on a stereoscopic parking garage provided by an embodiment of the present invention, the method comprises the following steps:

[0045] S10, receiving a start instruction for a target vehicle located in a preset three-dimensional parking garage area, wherein the start instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area.

[0046] The start command is a command automatically triggered by the user or the system through the system interface or remote control. It indicates a specific parking operation request, telling the system that the target vehicle is ready to enter the warehouse or needs to be parked in a specified location. This command usually contains the vehicle identification information of the target vehicle and the target parking space information where the target vehicle is located.

[0047] The preset three-dimensional parking garage area refers to the garage operation area preset by the system, which may contain multiple parking spaces and vehicle movement paths.

[0048] Specifically, after receiving the start command of the target vehicle in the preset three-dimensional parking garage area, the system will not directly perform the parking operation, but needs to go through steps S20-S60 to confirm that the target vehicle is in a non-abnormal state before the parking operation can be performed. Therefore, after receiving the start command, the detection system of the three-dimensional parking garage needs to ensure that there are no obstacles in the garage, the vehicle is in the correct position, and the garage personnel have left the preset three-dimensional parking garage area before the parking operation can be performed.

[0049] It can be seen that the detection system of the three-dimensional parking garage in this embodiment should respond quickly after receiving the start command to ensure the smoothness and efficiency of subsequent parking operations and reduce the waiting time of users.

[0050] S20. According to a preset image acquisition frequency, multi-directional image acquisition is performed on the target vehicle to obtain an image set of the target vehicle in multiple directions.

[0051] Among them, multi-directional image acquisition means capturing images of the target vehicle not only from a single angle but from multiple different angles, including the front, rear, sides and top of the vehicle. Through this comprehensive image acquisition method, the system can obtain more comprehensive vehicle information, thereby accurately analyzing and judging the vehicle's status, position and other important features.

[0052] Among them, multi-directional image acquisition usually involves multiple cameras, which are distributed in different locations in the garage to ensure that images of the target vehicle can be captured from different angles.

[0053] The preset image acquisition frequency refers to the fixed time interval or the number of frames that can be captured per second when the system acquires images. For example, if the preset image acquisition frequency is set to 30 frames per second, it means that the camera will capture 30 static images of the vehicle per second.

[0054] Among them, the preset image acquisition frequency can be set manually or according to the parameters of the camera module, and is not limited here.

[0055] Optionally, while presetting the image acquisition frequency, the image resolution will also be set to ensure that the clarity of the acquired images meets the requirements of subsequent image processing and analysis; and considering the changes in light in the garage, the camera module in the system should have the ability to automatically adjust exposure and focus to obtain high-quality images under different lighting conditions; in order to accurately analyze the movement and status of the vehicle, multiple cameras need to work synchronously to ensure that the acquired images are consistent in time; the acquired image sets need to be stored in the system and can be quickly transmitted to the processing unit for subsequent image preprocessing and detection.

[0056] It can be seen that the multi-directional image acquisition technology based on the preset image acquisition frequency in this embodiment can accurately obtain the state of the target vehicle and the movement trajectory in the garage, providing accurate and multi-directional image data for subsequent abnormality detection.

[0057] S30, performing image preprocessing on the image set of the target vehicle to obtain a preprocessed target image set of the target vehicle.

[0058] Among them, image preprocessing can improve image quality and enhance image features, thereby improving the accuracy and efficiency of subsequent processing. Image preprocessing may include but is not limited to removing interference information, enhancing useful features, standardizing image formats, etc.

[0059] According to different application scenarios and requirements, different image preprocessing techniques may need to be selected. For example, if the image has a lot of noise, denoising may be selected.

[0060] The target image set is an optimized image that is more suitable for target detection, recognition or other computer vision tasks. In this solution, the target image set includes pre-processed multi-directional images of the target vehicle, which together constitute a record of the target vehicle at different angles and at different or the same time points.

[0061] Furthermore, the preprocessed target image set of the target vehicle is used as input data for a processing module to perform target detection to identify and track the vehicle.

[0062] In one embodiment, the image preprocessing of the target vehicle's image set to obtain the preprocessed target image set of the target vehicle includes: performing denoising processing on the target vehicle's image set to obtain a denoised first image set of the target vehicle; performing data enhancement processing on the first image set of the target vehicle to obtain a processed second image set of the target vehicle; and performing correction processing on the second image set of the target vehicle to obtain a corrected target image set of the target vehicle.

[0063] The purpose of denoising is to remove irrelevant background noise from the image. Background noise may affect the accuracy of image analysis, especially in scenarios such as vehicle recognition and tracking, where clear images are essential. The first set of images obtained after denoising will be clearer than the original images.

[0064] Data augmentation increases the diversity of image data by applying a series of changes (such as rotation, scaling, cropping, etc.). This not only improves the generalization ability of the model, but also helps the system better identify vehicle images under different conditions. Through data augmentation, the first image set is transformed into a more diverse second image set.

[0065] The correction process includes correcting the distortion in the image, adjusting the direction and scale of the image, and ensuring that the image reflects the true situation of the vehicle. For the intelligent parking system, accurate vehicle images help improve the accuracy of vehicle positioning and recognition. The target image set obtained after correction reflects the cumulative effect of each processing step and provides high-quality image resources for subsequent tasks (such as vehicle recognition, classification, etc.).

[0066] It can be seen that in this embodiment, due to the influence of environmental factors (such as lighting changes, shadows, reflections, etc.) and equipment factors (such as camera resolution, lens distortion, etc.), the original captured image may have various defects. Through image preprocessing, these defects can be eliminated or reduced to improve the usability of the image.

[0067] S40: Perform image detection on the target image set to obtain a target object of each target image in the target image set.

[0068] The process of image detection can be model acquisition, image input, regional feature analysis, etc. Model acquisition can be to select or train a target detection model suitable for the current task; image input is to pass the preprocessed target image set as input data to the target detection model; regional feature analysis is to analyze the image according to the model and propose the area that may contain the target object (to be determined area), and extract features from the to be determined area, which are helpful to identify the target object.

[0069] Optionally, image detection includes not only identifying the existence of these target objects, but also determining their specific location, size and category in the image.

[0070] The target objects can be target vehicles, people, other objects, etc. Once the target objects are confirmed, their specific locations in the image will also be marked, which is crucial for subsequent tracking and behavior analysis.

[0071] In one embodiment, performing image detection on the target image set to obtain the target object of each target image in the target image set includes: acquiring a preset target detection model; inputting the target image set into the target detection model for image detection to obtain the to-be-determined area of ​​each target image in the target image set; performing regional feature analysis on the to-be-determined area of ​​each target image to obtain the target object of each target image in the target image set, wherein the target object is located in the to-be-determined area of ​​the corresponding target image.

[0072] Among them, the target detection model is usually based on deep learning algorithms, such as convolutional neural network (CNN), YOLO (You Only Look Once), SSD (Single Shot MultiBox Detector) or Faster R-CNN (Region-based Convolutional Neural Networks), etc. These models are trained with a large amount of data and can effectively identify and locate target objects in images.

[0073] Among them, the undetermined area refers to the image area that the target detection model initially determines may contain the target object. These areas need further analysis to confirm whether they really contain the target object.

[0074] The regional feature analysis may include but is not limited to the following: analyzing the geometric shape of the region, such as edges, corners, etc.; analyzing the texture pattern of the region, such as roughness, directionality, etc.; analyzing the color distribution and contrast within the region, etc.

[0075] Furthermore, by performing feature analysis on the area to be determined, the target object can be identified more accurately and false detection and missed detection can be reduced.

[0076] It can be seen that in this embodiment, the target object in the image is accurately identified and located through the target detection model and feature analysis, thereby reducing false detection and missed detection.

[0077] S50: Track the target object of each target image, and determine whether the target object of each target image is in an abnormal state.

[0078] Among them, target tracking algorithms are usually used to track target objects, such as Kalman filtering, particle filtering, deep learning tracking algorithms (such as Siamese networks), etc. These algorithms can track target objects in consecutive image frames in an image set and maintain tracking even when the target object moves or is partially occluded.

[0079] Among them, the abnormal state refers to the behavior or state of the target object in the stereo garage that does not conform to the normal operation mode. This may include but is not limited to the vehicle being parked in a non-designated area, a person being detected in the preset stereo garage area, a door not being closed, the trunk of the vehicle not being closed, etc.

[0080] For example, if the door is not completely closed when the vehicle is parked, it may affect the normal operation of the garage, such as the inability of the robotic arm to move the vehicle, or items falling from the vehicle during the movement, causing losses or safety hazards.

[0081] Optionally, methods for abnormal state detection may include, but are not limited to: analyzing the motion trajectory and behavior pattern of the target object and comparing them with preset normal behavior patterns; setting a series of rules or thresholds, and when certain parameters of the target object exceed these rules, the system determines it to be an abnormal state; the system monitors the state of the target object in real time, and immediately issues an alarm once an abnormality is detected.

[0082] Among them, detecting abnormal conditions is crucial to ensuring garage safety, preventing accidents, and responding to emergencies in a timely manner.

[0083] It can be seen that in this embodiment, the safety and efficiency of the garage are improved by tracking the target object and detecting abnormal conditions.

[0084] S60: When the target object of each target image is not in an abnormal state, the start instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area.

[0085] Among them, the above-mentioned starting instruction means that when the system confirms that the target vehicle and the target object in its image are not in an abnormal state, it is confirmed that the issued starting instruction can be executed, that is, the parking operation of the target vehicle in the preset multi-story parking garage area can be started.

[0086] Among them, the parking operation may include multiple steps, such as vehicle handling, positioning and fixing. These operations are automatically completed by the garage's mechanical system according to a preset program.

[0087] In one embodiment, after tracking the target object of each target image and determining whether the target object of each target image is in an abnormal state, the method further includes: when the target object of each target image is in an abnormal state, suspending the parking operation of the target vehicle in the preset stereoscopic parking garage area without issuing the start instruction.

[0088] Among them, if the target object is found to be in an abnormal state during the tracking and detection process, the system will not confirm the execution of the issued start command, but will take the following measures: the parking operation of the target vehicle will be suspended to prevent the risks that may be caused by continuing the operation in an abnormal state; an alarm may be triggered, and the garage manager or car owner may be notified of the abnormal state through a display screen, text message or other communication methods; an exception handling procedure may be initiated, such as guiding the car owner to readjust the vehicle position, or dispatching staff to the site for inspection and processing.

[0089] It can be seen that the detection system of the multi-story parking garage in this embodiment controls the process of parking operations by starting instructions while ensuring safety, and takes corresponding measures to ensure the normal operation of the garage and vehicle safety when an abnormal state is detected, further reducing the risk of accidents caused by human operational errors.

[0090] This method reduces the links of human judgment through automated image acquisition and processing, reduces errors caused by operator negligence, and improves the reliability of the system. Through continuous image acquisition and detection, the system can monitor the situation in the garage in real time, respond quickly to any abnormal situation, and take timely measures. Further, through real-time monitoring and tracking of the image of the target vehicle, the system can promptly detect abnormal conditions, such as vehicle position deviation, abnormal stop, etc., so as to avoid potential collisions and accidents and ensure the safety of vehicles and personnel in the garage. After confirming that the vehicle is safe, the system can perform parking operations through start instructions, reducing human intervention and improving the automation level and parking efficiency of the stereo garage. The system can be adjusted according to different garage structures and vehicle types, and has good adaptability and flexibility.

[0091] In one embodiment, tracking the target object of each target image and determining whether the target object of each target image is in an abnormal state includes: tracking the position of the target object of each target image to obtain a first path and a first position of the first path; analyzing whether the first path and the first position of the first path are within a preset position range; if not within the preset position range, determining that the target object of each target image is in an abnormal state.

[0092] Among them, position tracking uses image recognition technology, and the system continuously monitors the position of the target object in a series of continuous images (or video frames). By collecting this position information, the system can accurately understand the movement trajectory of the target object over time.

[0093] The first path refers to a trajectory formed by a series of position points of the target object within a period of time. These position points are usually obtained by analyzing continuous image frames in an image set.

[0094] The first position refers to the end point in the first path.

[0095] The preset position range may be different according to the type of the target object, and is not limited here. For example, when the target object is a door, the preset position range is the position range of the door when it is closed, which is determined according to the vehicle model.

[0096] After having the category information of the target object and the corresponding preset position range, the system can determine the position of the target object according to the actual situation to see whether it is within the preset range. If the position of the target object deviates from the preset range, it may be judged as an abnormal state.

[0097] For example, taking a car door as an example, the position range of the door when closed may be different for different vehicle models. The monitoring system should be able to identify the appropriate preset position range according to the vehicle model, and then determine whether the door is closed normally. If the door is in an abnormal state (for example, not closed), the system can promptly remind relevant personnel to check and handle it.

[0098] It can be seen that in this embodiment, by acquiring and analyzing the target data of the target object, its position is predicted, and further accurate abnormal state determination is performed.

[0099] In one embodiment, after tracking the position of the target object of each target image to obtain a first path and a first position of the first path, the method further includes: acquiring target data corresponding to the target object of each target image; predicting the position of the target object of each target image based on the target data corresponding to the target object of each target image to obtain a second path; judging whether a path abnormality occurs based on the target data corresponding to the target object of each target image, the first path, and the second path; if a path abnormality occurs, performing data analysis based on the target data corresponding to the target object of each target image to determine whether the target object of each target image is in an abnormal state.

[0100] The target data may be a target location, that is, location information of a target object.

[0101] Among them, location prediction can learn the movement pattern of the object based on machine learning or deep learning algorithms, combined with target data and historical data, and predict its future path accordingly. Location prediction is to predict the possible location of the target object in the future, which helps to detect potential path anomalies and risks in advance.

[0102] For example, position prediction needs to take into account the specific behavior and movement patterns of the target object. For example, the position prediction of a car door needs to take into account the opening and closing action and speed of the door, not just simple linear motion.

[0103] The second path is the future path predicted based on the target data, which is usually predicted through time series analysis, machine learning models or physical models.

[0104] Specifically, path anomaly refers to a significant difference between the actual path (first path) and the predicted path (second path), which may indicate that the actual movement of the target object does not conform to the expected pattern; judging path anomaly usually involves comparing the current actual position with the predicted position, and analyzing the continuity and smoothness of the path.

[0105] Specifically, when a path anomaly is detected, the system will further analyze the target data to determine whether there is an abnormal state. This may include in-depth analysis of the target object's movement pattern, speed change, direction change, etc. If the data analysis shows that the target object's movement does not conform to the normal behavior pattern, the system will determine that the target object is in an abnormal state.

[0106] For example, if a door does not close properly when it should, the system may determine that the door is in an abnormal state.

[0107] It can be seen that in this embodiment, by predicting the future position of the target object, the system can actively monitor and prevent potential safety risks. The detection of abnormal conditions is not only based on the current position information, but also includes a comprehensive analysis of the historical and predicted data of the target object.

[0108] In one embodiment, judging whether a path abnormality occurs based on the target data corresponding to the target object of each target image, the first path, and the second path includes: determining a preset distance threshold corresponding to the target object of each target image based on the target data corresponding to the target object of each target image; comparing a first position of the first path and a second position of the second path; and judging that a path abnormality occurs when the distance between the first position and the second position does not meet the corresponding preset distance threshold.

[0109] The preset distance threshold is a safety range set according to the expected movement behavior of the target object in the stereo garage. This threshold represents the maximum acceptable distance difference between the position of the target object and the predicted position during normal movement. The preset distance threshold can be obtained according to the category of the target object, and the preset distance threshold can be set manually or not, which is not limited here.

[0110] The target data includes information such as the type, size, and motion characteristics of the target object, which is used to calculate and determine a preset distance threshold suitable for the object. For example, for a vehicle door, the threshold may take into account the amplitude and speed of the door opening and closing.

[0111] Specifically, the first path is usually calculated based on the actual motion trajectory of the target object, and the second path is the future position or expected trajectory predicted based on the target data; the two paths are compared to ensure that the actual motion of the target object is consistent with the expectation.

[0112] If the distance between the first position (actual position) and the second position (predicted position) exceeds the preset distance threshold, the system will consider that the target object's behavior is not as expected, and thus determine that a path anomaly has occurred. A path anomaly may indicate that the target object has encountered an obstacle, an operation error, or other unexpected situations.

[0113] It can be seen that in this embodiment, by comparing the actual and predicted path positions to determine whether the target object deviates from the normal motion trajectory, corresponding measures can be taken in time when an abnormality is detected, which helps to ensure the safe and efficient operation of the stereo garage.

[0114] In one embodiment, the target data includes a target category. If a path abnormality occurs, data analysis is performed based on the target data corresponding to the target object of each target image to determine whether the target object of each target image is in an abnormal state, including: determining a first dwell time based on the target category corresponding to the target object of each target image; if the first dwell time exceeds the preset dwell time corresponding to the corresponding target category, determining that the target object of each target image is in an abnormal state.

[0115] The target category refers to the type of target object in the garage, such as vehicles, pedestrians, handling equipment, etc. Different types of target objects have different behavior patterns in the garage, so their residence time standards will also be different. The classification process is usually completed by computer vision and deep learning algorithms, which can understand the content in the image and accurately classify the objects.

[0116] The first dwell time refers to the length of time that the target object stays at a certain location in the garage. This time is usually calculated by analyzing consecutive image frames in the image set, that is, the time from when the target object is first detected at a certain location to when it leaves the location.

[0117] The preset dwell time is a time threshold set for different target categories, which reflects the reasonable time range for objects of this category to stay at a certain location under normal circumstances. The preset dwell time is determined based on historical data analysis, safety considerations, garage operation efficiency and other factors, and is not a unique limitation here.

[0118] If the actual stay time (first stay time) of the target object exceeds the preset stay time set for its category, the system will consider this to be an abnormal situation. An abnormal stay time may indicate that the target object has encountered a problem. For example, after receiving the start command, the system detects that person A has been in the car, and determines that it is in an abnormal state at this time.

[0119] Furthermore, when an abnormal residence time of a target object is detected, the system will mark it as an abnormal state and may take further actions, such as sounding an alarm, notifying management personnel, and initiating emergency response procedures.

[0120] It can be seen that in this embodiment, by identifying the target category through the target image and analyzing its residence time in a specific area, it is possible to effectively determine whether the object is in an abnormal state, thereby improving the accuracy of the detection method based on the stereoscopic parking garage.

[0121] It should be noted that, in each of the above-mentioned embodiments, there is not necessarily a certain order between the above-mentioned steps. A person skilled in the art can understand, based on the description of the embodiments of the present application, that in different embodiments, the above-mentioned steps may have different execution orders, that is, they may be executed in parallel, may be executed interchangeably, and so on.

[0122] As another aspect of the embodiment of the present application, the embodiment of the present application provides a detection device based on a stereoscopic parking garage. The detection device based on the stereoscopic parking garage can be a software module, and the software module includes a number of instructions, which are stored in a memory, and the processor can access the memory and call the instructions for execution to complete the detection method based on the stereoscopic parking garage described in the above various embodiments.

[0123] See also Figure 3 , Figure 3 Schematic diagram of a detection device based on a three-dimensional parking garage provided in an embodiment of the present application. Figure 3 As shown, a detection system applied to a stereo garage, the detection device 300 based on the stereo garage includes:

[0124] A receiving unit, configured to receive a confirmation instruction of a target vehicle located in a preset three-dimensional parking garage area, wherein the confirmation instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area;

[0125] The acquisition unit 301 is used to acquire images of the target vehicle according to a preset image acquisition frequency to obtain an image set of the target vehicle;

[0126] The processing unit 302 is used to perform image preprocessing on the image set of the target vehicle to obtain a preprocessed target image set of the target vehicle;

[0127] A detection unit 303 is used to perform image detection on the target image set to obtain a target object of each target image in the target image set;

[0128] A determination unit 304, configured to track a target object in each target image and determine whether the target object in each target image is in an abnormal state;

[0129] The determining unit 304 is further configured to, when the target object of each target image is not in an abnormal state, perform a parking operation on the target vehicle in the preset three-dimensional parking garage area through the start instruction.

[0130] This method reduces the links of human judgment through automated image acquisition and processing, reduces errors caused by operator negligence, and improves the reliability of the system. Through continuous image acquisition and detection, the system can monitor the situation in the garage in real time, respond quickly to any abnormal situation, and take timely measures. Further, through real-time monitoring and tracking of the image of the target vehicle, the system can promptly detect abnormal conditions, such as vehicle position deviation, abnormal stop, etc., so as to avoid potential collisions and accidents and ensure the safety of vehicles and personnel in the garage. After confirming that the vehicle is safe, the system can perform parking operations through start instructions, reducing human intervention and improving the automation level and parking efficiency of the stereo garage. The system can be adjusted according to different garage structures and vehicle types, and has good adaptability and flexibility.

[0131] In one embodiment, in tracking the target object of each target image and determining whether the target object of each target image is in an abnormal state, the determination unit 304 is further used to: perform position tracking on the target object of each target image to obtain a first path and a first position of the first path; analyze whether the first path and the first position of the first path are within a preset position range; if not within the preset position range, determine that the target object of each target image is in an abnormal state.

[0132] In one embodiment, after the position of the target object of each target image is tracked to obtain a first path and a first position of the first path, the determination unit 304 is further used to: obtain target data corresponding to the target object of each target image; predict the position of the target object of each target image according to the target data corresponding to the target object of each target image to obtain a second path; determine whether a path abnormality occurs according to the target data corresponding to the target object of each target image, the first path, and the second path; if a path abnormality occurs, perform data analysis according to the target data corresponding to the target object of each target image to determine whether the target object of each target image is in an abnormal state.

[0133] In one embodiment, in determining whether a path abnormality occurs based on the target data corresponding to the target object of each target image, the first path, and the second path, the determination unit 304 is further used to: determine a preset distance threshold corresponding to the target object of each target image based on the target data corresponding to the target object of each target image; compare the first position of the first path and the second position of the second path; and determine that a path abnormality occurs when the distance between the first position and the second position does not meet the corresponding preset distance threshold.

[0134] In one embodiment, if a path abnormality occurs, data analysis is performed based on the target data corresponding to the target object of each target image to determine whether the target object of each target image is in an abnormal state, and the determination unit 304 is also used to: determine a first stay time based on the target category corresponding to the target object of each target image; if the first stay time exceeds the preset stay time corresponding to the corresponding target category, it is determined that the target object of each target image is in an abnormal state.

[0135] In one embodiment, after tracking the target object of each target image and determining whether the target object of each target image is in an abnormal state, the determination unit 304 is further used to: when the target object of each target image is in an abnormal state, suspend the parking operation of the target vehicle in the preset stereoscopic parking garage area without issuing the start instruction.

[0136] In one embodiment, in performing image preprocessing on the image set of the target vehicle to obtain the preprocessed target image set of the target vehicle, the processing unit 302 is further used to: perform denoising on the image set of the target vehicle to obtain a denoised first image set of the target vehicle; perform data enhancement on the first image set of the target vehicle to obtain a processed second image set of the target vehicle; and perform correction on the second image set of the target vehicle to obtain a corrected target image set of the target vehicle.

[0137] In one embodiment, in performing image detection on the target image set to obtain the target object of each target image in the target image set, the detection unit 303 is further used to: obtain a preset target detection model; input the target image set into the target detection model to perform image detection to obtain the area to be determined for each target image in the target image set; perform regional feature analysis on the area to be determined for each target image to obtain the target object of each target image in the target image set, wherein the target object is located in the area to be determined for the corresponding target image.

[0138] It should be noted that the above-mentioned detection device based on a three-dimensional parking garage can execute the detection method based on a three-dimensional parking garage provided in the embodiment of this application, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not fully described in the embodiment of the detection device based on a three-dimensional parking garage, please refer to the detection method based on a three-dimensional parking garage provided in the embodiment of this application.

[0139] See also Figure 4 , Figure 44 is a schematic diagram of a computer device provided in an embodiment of the present application. The computer device includes one or more processors 41 and a memory 42. The memory 42 is connected to the one or more processors 41, for example, connected to the processor 41 via a bus.

[0140] The processor 41 is configured to support the computer device to perform the corresponding functions in the method in the above method embodiment. The processor 41 can be a central processing unit (CPU), a network processor (NP), a hardware chip or any combination thereof. The above hardware chip can be an application specific integrated circuit (ASIC), a programmable logic device (PLD) or a combination thereof. The above PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL) or any combination thereof.

[0141] The memory 42 is used to store program codes, etc. The memory 42 may include a volatile memory (VM), such as a random access memory (RAM); the memory 42 may also include a non-volatile memory (NVM), such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD); the memory 42 may also include a combination of the above-mentioned types of memories.

[0142] The memory 42 can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the detection method based on the three-dimensional parking garage in the embodiment of the present application. The processor 41 executes various functional applications and data processing of the detection method based on the three-dimensional parking garage and the detection device based on the three-dimensional parking garage by running the non-volatile software programs, instructions and modules stored in the memory 42, that is, realizes the functions of each module or unit of the detection method based on the three-dimensional parking garage and the detection device based on the three-dimensional parking garage provided in the above method embodiment.

[0143] The memory 42 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function. The data storage area may store data created based on the use of the detection device based on the stereo garage, etc. In some embodiments, the memory 42 may optionally include a memory 42 remotely arranged relative to the processor 41, and these remote memories 42 may be connected to the detection device based on the stereo garage via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0144] The one or more modules are stored in the memory 42. When executed by the one or more processors 41, the detection method based on the stereoscopic parking garage in any of the above-mentioned method embodiments is executed, for example, the method steps described in the above-mentioned method embodiments are executed to realize the functions of the modules described in the above-mentioned device embodiments.

[0145] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a computer, the computer executes the method described in the above embodiment.

[0146] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0147] The above disclosure is only the preferred embodiment of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A detection method based on a stereo garage, characterized in that: A detection system applied to a stereo garage, the method comprising: Receiving a start instruction for a target vehicle in a preset three-dimensional parking garage area, the start instruction being used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area; According to a preset image acquisition frequency, multi-directional image acquisition is performed on the target vehicle to obtain an image set of the target vehicle in multiple directions; Performing image preprocessing on the image set of the target vehicle to obtain a preprocessed target image set of the target vehicle; Performing image detection on the target image set to obtain a target object of each target image in the target image set; Tracking the target object of each target image, and determining whether the target object of each target image is in an abnormal state; When the target object of each target image is not in an abnormal state, the start instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area.

2. The method according to claim 1, characterized in that: Tracking the target object of each target image and determining whether the target object of each target image is in an abnormal state includes: Tracking the position of the target object in each target image to obtain a first path and a first position of the first path; analyzing whether the first path and the first position of the first path are within a preset position range; If it is not located in the preset position range, it is determined that the target object of each target image is in an abnormal state.

3. The method according to claim 2, characterized in that After tracking the position of the target object in each target image to obtain the first path and the first position of the first path, the method further includes: Acquire target data corresponding to the target object of each target image; Predicting the position of the target object of each target image according to the target data corresponding to the target object of each target image to obtain a second path; determining whether a path abnormality occurs according to the target data corresponding to the target object of each target image, the first path, and the second path; If a path abnormality occurs, data analysis is performed based on the target data corresponding to the target object of each target image to determine whether the target object of each target image is in an abnormal state.

4. The method according to claim 3, characterized in that: The determining whether a path abnormality occurs according to the target data corresponding to the target object of each target image, the first path, and the second path includes: Determining a preset distance threshold corresponding to the target object of each target image according to the target data corresponding to the target object of each target image; comparing a first position of the first path and a second position of the second path; When the distance between the first position and the second position does not meet the corresponding preset distance threshold, it is determined that a path abnormality occurs.

5. The method according to claim 3, characterized in that: The target data includes a target category. If a path abnormality occurs, data analysis is performed according to the target data corresponding to the target object of each target image to determine whether the target object of each target image is in an abnormal state, including: Determining a first dwell time according to a target category corresponding to the target object of each target image; If the first stay time exceeds the preset stay time corresponding to the corresponding target category, it is determined that the target object of each target image is in an abnormal state.

6. The method according to claim 1, characterized in that After tracking the target object of each target image and determining whether the target object of each target image is in an abnormal state, the method further includes: When the target object of each target image is in an abnormal state, the parking operation of the target vehicle in the preset three-dimensional parking garage area is suspended without passing the start instruction.

7. The method according to claim 1, characterized in that The performing image preprocessing on the image set of the target vehicle to obtain the preprocessed target image set of the target vehicle includes: Performing denoising on the image set of the target vehicle to obtain a first image set of the target vehicle after denoising; Performing data enhancement processing on the first image set of the target vehicle to obtain a processed second image set of the target vehicle; The second image set of the target vehicle is corrected to obtain a corrected target image set of the target vehicle.

8. The method according to claim 1, characterized in that The performing image detection on the target image set to obtain the target object of each target image in the target image set includes: Get the preset target detection model; Inputting the target image set into the target detection model to perform image detection, and obtaining a region to be determined for each target image in the target image set; Performing regional feature analysis on the to-be-determined region of each target image to obtain a target object of each target image in the target image set, wherein the target object is located in the to-be-determined region of the corresponding target image.

9. A detection device based on a three-dimensional parking garage, applied to a detection system of a three-dimensional parking garage, the device comprising: A receiving unit, configured to receive a confirmation instruction of a target vehicle located in a preset three-dimensional parking garage area, wherein the confirmation instruction is used to perform a parking operation on the target vehicle in the preset three-dimensional parking garage area; An acquisition unit, used to acquire images of the target vehicle according to a preset image acquisition frequency to obtain an image set of the target vehicle; A processing unit, used for performing image preprocessing on the image set of the target vehicle to obtain a preprocessed target image set of the target vehicle; A detection unit, configured to perform image detection on the target image set to obtain a target object of each target image in the target image set; a determination unit, configured to track a target object in each target image and determine whether the target object in each target image is in an abnormal state; The determining unit is further configured to, when the target object of each target image is not in an abnormal state, perform a parking operation on the target vehicle in the preset three-dimensional parking garage area through the start instruction.

10. A computer device, comprising a memory and a processor, wherein the memory is connected to the processor, and the processor is used to execute one or more computer programs stored in the memory, and when the processor executes the one or more computer programs, the computer device implements the method according to any one of claims 1 to 8.

Images