A proprietary parking space autonomous management device based on target detection
Through the combination of YOLOv5 deep learning algorithm and image acquisition module, the accurate detection and identification of license plates are achieved, the problem of unauthorized vehicles occupying proprietary parking spaces is solved, and the efficiency and accuracy of parking space management are improved.
Patent Information
- Application Number
- CN202410354399.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-27
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-03-27
AI Technical Summary
The prior art is difficult to improve the accuracy of license plate detection and character recognition under various environmental conditions, and cannot effectively prevent unauthorized vehicles from occupying proprietary parking spaces, and lacks intelligent management methods.
The object detection technology based on YOLOv5 deep learning algorithm is adopted, combined with image acquisition, character recognition and communication modules, real-time detection and recognition of license plates are realized, and unauthorized vehicles are warned through capture and SMS/mail notification mechanisms, and a basis for management is provided.
It improves the accuracy and efficiency of license plate inspection, prevents unauthorized vehicles from occupying proprietary parking spaces, reduces the dispute rate, and improves the utilization rate and management efficiency of parking spaces.
Smart Images

Figure CN118736785B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of autonomous management of dedicated parking spaces, and in particular to an autonomous management device and method for dedicated parking spaces based on target detection. Background Art
[0002] With the development of society and the improvement of people's lives, cars have become the preferred means of transportation for people. According to statistics, the number of motor vehicles in China will reach 435 million in 2023, including 336 million cars. However, with the sharp increase in the number of cars, the contradiction between supply and demand of parking spaces has become increasingly prominent, and the phenomenon of illegal parking is prevalent, which has brought many inconveniences to urban traffic. Especially in some residential areas, office areas and other places, the management of private parking spaces has become a prominent problem. Cars occupy them at will, causing waste of resources and affecting the normal lives of residents.
[0003] In order to solve the parking space management problem, it is necessary to use scientific and technological means to manage parking spaces intelligently, such as license plate detection and recognition systems. As a widely used machine vision technology in traffic scene recognition, license plate detection and recognition systems can realize the automatic collection and recognition of vehicle information, and also provide effective technical support for parking space management. Through license plate recognition technology, we can realize the automatic recording of vehicle entry and exit and the autonomous management of parking spaces, thereby improving the efficiency of parking space management and alleviating the problem of parking difficulties. However, the current intelligent management of parking spaces is mainly aimed at areas such as garages where a large number of vehicles are parked or passing through, and there is less intelligent management of dedicated parking spaces. Therefore, it is very worthwhile to develop a dedicated parking space autonomous management device based on target detection. Summary of the invention
[0004] The main technical problems solved by the present invention are: improving the accuracy of license plate detection and character recognition under various environmental conditions; preventing dedicated parking spaces from being occupied by unauthorized vehicles by utilizing a warning and expulsion mechanism; capturing unauthorized vehicles occupying dedicated parking spaces, notifying users via text messages or emails to remind them that their parking spaces are occupied and their warnings have been ignored, and providing a basis for subsequent management and punishment.
[0005] In order to solve the above problems, the technical solution of the present invention proposes a dedicated parking space autonomous management device and management method based on target detection.
[0006] To achieve the above object, the present invention adopts the following technical solution:
[0007] A proprietary parking space autonomous management device based on target detection, characterized in that it includes a parking space terminal and a user terminal;
[0008] The parking terminal includes an image acquisition device, a voice device and a control unit. The control unit includes a communication module, a storage module, an image target detection module, an image character recognition module, a management module, a capture module, and a text-to-speech module. The communication module, the storage module, the image target detection module, the image character recognition module, the capture module, and the text-to-speech module are all connected to the management module;
[0009] The communication module is used for communication, including data communication and image communication; the communication module includes a WIFI submodule and a cellular communication submodule, which are used to be managed by the management module to connect to the network and send vehicle-related information to the user terminal, and the vehicle-related information includes an image captured by the image acquisition device controlled by the capture module and the time of the capture of the vehicle occupying the parking space;
[0010] The storage module is used to store vehicle information, which includes preset license plate characters, license plate characters obtained by the image character recognition module, and the time when the license plate characters are obtained by the image character recognition module;
[0011] The image target detection module is used to perform image detection on the continuous images collected by the image acquisition device, and the image detection includes performing target detection on the license plate on the image containing the license plate;
[0012] The image character recognition module is used to perform image character recognition, including performing character recognition on the license plate detected by the image target detection module to obtain the license plate characters;
[0013] The management module is used to manage the work of the modules connected thereto, and also includes license plate character comparison, which is to compare the license plate characters obtained by the image character recognition module with the preset license plate characters. If the license plate characters recognized by the character comparison are not matched with the preset license plate characters for multiple consecutive times, the image acquisition device is triggered to capture the image of the vehicle occupying the parking space to form vehicle-related information, which includes the captured image and the time of the capture, and the vehicle-related information is stored in the storage module, and the vehicle-related information is sent to the user terminal through the communication module network.
[0014] The capture module is used to control the image acquisition device to capture the vehicle occupying the parking space;
[0015] The text-to-speech module is used to perform text-to-speech operations, including calling the control module according to the result of the license plate character comparison performed by the management module, and controlling the voice device to issue a prompt message.
[0016] The user terminal includes a user terminal device and is configured with a short message service and an email service for receiving vehicle-related information, including data and images, sent from a communication module.
[0017] The target detection-based dedicated parking space autonomous management device is characterized in that the image target detection module adopts the YOLOv5 deep learning algorithm. In the present invention, the YOLOv5 algorithm extracts features from the input image through a convolutional neural network, and uses an anchor frame mechanism to predict the position, size and category of the license plate to achieve target detection of the license plate.
[0018] A dedicated parking space autonomous management device based on target detection, a management method for dedicated parking space management using the dedicated parking space autonomous management device based on target detection, the specific steps are as follows:
[0019] (1) Image acquisition: The parking area is continuously monitored in real time through image acquisition devices such as high-definition cameras to obtain a continuous image data stream.
[0020] (2) Image preprocessing: Perform preprocessing operations such as grayscale conversion, noise reduction, and enhancement on continuously acquired images to improve the accuracy and efficiency of subsequent processing.
[0021] (3) Object detection: The YOLOv5 deep learning algorithm is used to detect vehicles and license plates in the preprocessed images.
[0022] The YOLOv5 algorithm quickly and accurately identifies the vehicle and its license plate location in the image and outputs the bounding box coordinates of the license plate.
[0023] (4) Character recognition: Based on the license plate bounding box coordinates obtained by target detection, the license plate image within the bounding box is subjected to license plate character recognition, and the character information in the image is converted into text data using optical character recognition (OCR) technology.
[0024] (5) Character comparison: The recognized license plate characters are compared with the preset dedicated parking license plate character library to determine whether it is an authorized vehicle.
[0025] (6) Prompt information generation and voice prompt: Generate corresponding prompt information based on the character comparison results. If it is an authorized vehicle, a prompt message such as "Welcome home" is generated; if it is an unauthorized vehicle, a prompt message such as "Do not park in the dedicated parking space, please leave immediately" is generated. Text-to-speech (TTS) technology is used to play the generated prompt information through a voice device.
[0026] The dedicated parking space autonomous management device and management method based on target detection described in the present invention have the following advantages:
[0027] It can effectively detect license plate characters in images containing license plates, realize autonomous management of dedicated parking spaces, prevent dedicated parking spaces from being occupied by unauthorized vehicles, and reduce the probability of unnecessary troubles and disputes due to parking space occupation. In addition, it can automatically identify vehicles and realize the management of multiple license plates corresponding to one parking space, thereby improving parking efficiency and the utilization rate of dedicated parking spaces. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0029] Figure 1 A distribution diagram of the dedicated parking space autonomous management device based on target detection of the present invention;
[0030] Figure 2 This is a diagram showing the license plate recognition result display interface of the proprietary parking space autonomous management device based on target detection of the present invention;
[0031] Figure 3 The figure is a flow chart of the use of the dedicated parking space autonomous management device based on target detection of the present invention. DETAILED DESCRIPTION
[0032] Combination Figure 1 The content shown, this embodiment discloses a dedicated parking space autonomous management device based on target detection, including a parking space terminal and a user terminal connected to each other. The parking space terminal includes an image acquisition device and a control unit, and the image acquisition device is connected to the control unit;
[0033] The control unit includes a communication module, a storage module, an image target detection module, an image character recognition module, a management module, a capture module, and a text-to-speech module, and the communication module, the storage module, the image target detection module, the image character recognition module, the capture module, and the text-to-speech module are all connected to the management module;
[0034] The communication module is used for communication, including networking and calling the capture module, and sending vehicle-related information to the user terminal, wherein the vehicle-related information includes the image captured by the capture module to control the image acquisition device to capture the vehicle occupying the parking space and the capture time, wherein the communication module includes a WIFI submodule and a cellular communication submodule to achieve networking;
[0035] The storage module is used to store vehicle information, which includes preset license plate characters, license plate characters obtained by the image character recognition module, and the time when the license plate characters are obtained by the image character recognition module;
[0036] The image target detection module is used to perform image detection on the continuous images collected by the image acquisition device, and the image detection includes performing target detection on the license plate on the image containing the license plate;
[0037] The image character recognition module is used to perform image character recognition, including performing character recognition on the license plate detected by the image target detection module to obtain the license plate characters;
[0038] The management module is used to manage the work of the modules connected thereto, and also includes license plate character comparison, where the license plate character comparison is to compare the license plate characters obtained by the image character recognition module with the preset license plate characters;
[0039] The capture module is used to control the image acquisition device to capture the vehicle occupying the parking space.
[0040] The dedicated parking space autonomous management device based on target detection also includes a voice device, and the parking space terminal also includes a text-to-speech module and a control module connected to the management module. The text-to-speech module is used to perform text-to-speech operations, including calling the control module according to the result of the license plate character comparison by the management module, and controlling the voice device to issue a prompt message through the control module;
[0041] The voice device is associated with the capture module.
[0042] The above communication module includes data communication and image communication.
[0043] The above user terminal is equipped with SMS service and mailbox service.
[0044] Combination Figure 2 The content shown in this embodiment describes a dedicated parking space autonomous management device based on target detection, including a license plate recognition result display interface of the dedicated parking space autonomous management device based on target detection. The interface can display the recognized license plate information in real time and compare it with the preset valid license plate characters to determine whether it is an authorized vehicle.
[0045] Combination Figure 3The content shown in the embodiment describes a dedicated parking space autonomous management device based on target detection, which uses the above-mentioned dedicated parking space autonomous management device based on target detection to manage dedicated parking spaces, and pre-sets the number of valid license plate characters, the first time length, the second time length, the third time length, the fourth time length, and the fifth time length. For example, the pre-set number of valid license plate characters, the first time length, the second time length, the third time length, the fourth time length, and the fifth time length are set to 5, 5s, 15s, 30s, 45s, and 60s respectively. In addition, before the management method is executed, one or more preset license plate characters are pre-stored in the application corresponding to the management method. The specific management method includes the following steps:
[0046] Step 1: The image acquisition device continuously acquires images of the parking space and transmits the images to the image target detection module for target detection of the license plate:
[0047] (1) If the detection result is valid, that is, the license plate is detected, go to step 2;
[0048] (2) If the detection result is invalid, that is, the license plate is not detected, go to step 1.
[0049] Step 2: The image character recognition module performs character recognition on the license plate for the valid detection results.
[0050] If the character recognition result is valid, that is, the number of recognized valid license plate characters is greater than or equal to the preset number of valid license plate characters, then proceed to step 3;
[0051] If the character recognition result is valid, that is, the number of recognized valid license plate characters is less than the preset number of valid license plate characters, then go to step 1.
[0052] Step 3: The management module compares the license plate characters obtained by the image character recognition module with the preset license plate characters:
[0053] (1) If the characters on the license plate match the preset characters, the text-to-language module calls the control module to control the voice device to emit a corresponding welcome language, and then proceeds to step 4;
[0054] (2) If the characters of the license plate do not match the preset characters, the text-to-language module calls the control module to control the voice device to issue a corresponding warning language, and then goes to step 5.
[0055] Step 4: The management module continuously tracks the last recognition time of the image character recognition of the specific license plate. When the specific license plate cannot be recognized again after exceeding a preset first time length, the text-to-speech module calls the control module to control the voice device to emit a corresponding farewell language and enter step 1.
[0056] Step 5: The management module and the management storage module respectively continuously track the first and last recognition times of the image character recognition of the specific license plate, and proceed to step 5.1;
[0057] Step 5.1: When the last recognition time minus the first recognition time exceeds the preset second time length, the text-to-speech module calls the control module to control the voice device to issue a corresponding first warning language, and when the last recognition time minus the first recognition time is between the second time length and the preset third time length:
[0058] (1) If the vehicle leaves, the management module continuously tracks the last recognition time of the image character recognition of the specific license plate. When the specific license plate cannot be recognized again after a first time length, the text-to-speech module calls the control module to control the voice device to emit a corresponding thank you cooperation language, and then enters step 1;
[0059] (2) If the vehicle has not left, that is, the last recognition time minus the first recognition time exceeds the third time length, go to step 5.2.
[0060] Step 5.2: When the last recognition time minus the first recognition time exceeds the third time length, the text-to-speech module calls the control module to control the voice device to issue a corresponding second warning language, and when the last recognition time minus the first recognition time is between the third time length and the preset fourth time length:
[0061] (1) If the vehicle leaves, the management module and the storage module continue to track the last recognition time of the image character recognition of the specific license plate. When the specific license plate cannot be recognized again after a first time length, the text-to-language module calls the control module to control the voice device to emit a corresponding thank you cooperation language, and then enters step 1;
[0062] (2) If the vehicle has not left, that is, the last recognition time minus the first recognition time exceeds the fourth time length, go to step 5.3.
[0063] Step 5.3: When the last recognition time minus the first recognition time exceeds the fourth time length, the text-to-speech module calls the control module to control the voice device to issue a corresponding third or final warning language when the last recognition time minus the first recognition time is between the fourth time length and the preset fifth time length:
[0064] (1) If the vehicle leaves, the management module and the storage module continue to track the last recognition time of the image character recognition of the specific license plate. When the specific license plate cannot be recognized again after a first time length, the text-to-language module calls the control module to control the voice device to emit a corresponding thank you cooperation language, and then enters step 1;
[0065] (2) If the vehicle has not left, that is, the last recognition time minus the first recognition time exceeds the fifth time length, go to step 6.
[0066] Step 6, the management module manages the illegal parking capture module, calls the image acquisition device to capture images of illegally parked vehicles, and manages the text-to-speech module, calls the control module, controls the voice device to issue the corresponding illegal parking notification language, and manages the communication module to notify the user terminal of the captured vehicle images and information via SMS and email, and manages the storage module to store illegal parking images and illegal parking data.
[0067] In step 1, the image target detection module can use the YOLOv5 deep learning algorithm to detect the target of the license plate. When using the YOLOv5 deep learning algorithm to detect the target of the license plate, the target detection network of the YOLOv5 version with a light algorithm and fast response is preferred. The principle and process of YOLO:
[0068] When the target image (the image passed into the image object detection module) is input into the YOLO network, it will be divided into S×S grids of the same size by the model, and different grids will be responsible for different category information. If these grids detect the target, they will be responsible for detecting the target. These detections include the current target position, size and probability of belonging to the category.
[0069] The workflow of the YOLO model is divided into the following three stages:
[0070] (1) Feature extraction: The backbone network of YOLO performs feature extraction. The design of the YOLO backbone network refers to the network structure of GooLeNet, but does not use the inception module in GooLeNet. Instead, it uses 1×1 convolutional layers and 3×3 convolutional layers as replacements.
[0071] (2) Bounding box generation and regression. Its network structure contains 4 convolutional layers and 2 MLP layers. After inputting the features extracted in the previous text, an S×S×30 output is obtained;
[0072] (3) Bounding box screening: The target box screening of the YOLO model will be performed through the non-maximum suppression algorithm (NMS). The principle is that the suppression algorithm in the model will calculate and discard the bounding boxes with scores below the threshold, and use the NMS algorithm to remove redundant bounding boxes to obtain the best bounding box to represent the location, size and category of the detected target. The operation of repeated iterative traversal and bounding box elimination will be repeated in the above process until the best bounding box is obtained to achieve the goal.
[0073] The target detection algorithm can use numerical values to express the performance information of the algorithm. The present invention uses recall, precision, mean average precision (mAP) and FPS as the model evaluation indicators of the experimental results of the license plate target detection model used by the above target detection module. The expressions of recall and precision are as follows:
[0074]
[0075]
[0076] Among them, TP (True Positive) is the number of samples that are both predicted and actual positive, FP (False Positive) is the number of samples that are both predicted and actual negative, and FN (False Negative) is the number of samples that are predicted to be negative but are actually positive.
[0077] mAP (mean Average Precision) is an important indicator for evaluating model performance in deep learning models. It is the average AP of all categories. AP (Average Precision) represents the average accuracy under different recall rates. The mathematical expression of mAP is as follows:
[0078]
[0079]
[0080] Among them, n represents the category and N represents the total number of categories.
[0081] At the application level, an excellent target detection algorithm should not only ensure accuracy and precision, but also speed. This article uses FPS (Frames Per Second) to represent the number of images processed by the detector per second to represent the detection speed.
[0082]
[0083] Among them, t is the time to detect a picture, in seconds.
[0084] In the image object detection module, the CCPD dataset and YOLOv5s algorithm are used to train the license plate object detection model, and the final results are analyzed. After 100 rounds of training, the positioning loss and target confidence loss are both less than 0.02, and the mAP, accuracy and recall rate tend to be stable as the rounds increase. Finally, the model accuracy is 99.36%, the recall rate reaches 100%, the mAP@0.5 is 0.99, and the FPS is 39.52.
[0085] Result description:
[0086] The YOLOv5 deep learning algorithm has a very high model accuracy. Compared with other algorithms (UNET - 91%), it has better accuracy, stability and speed for license plate target detection.
[0087] The parts of the present invention that are not described in detail are common knowledge to those skilled in the art. The above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the protection scope of the present invention.
Claims
1. A proprietary parking space autonomous management device based on target detection, It is characterized in that Including parking space terminal and user terminal, The parking terminal includes an image acquisition device, a voice device and a control unit. The control unit includes a communication module, a storage module, an image target detection module, an image character recognition module, a management module, a capture module, and a text-to-speech module. The communication module, the storage module, the image target detection module, the image character recognition module, the capture module, and the text-to-speech module are all connected to the management module. (1) The communication module is used for communication, including data communication and image communication; the communication module includes a WIFI submodule and a cellular communication submodule, which are used to be managed by the management module to connect to the network and send vehicle-related information to the user terminal, and the vehicle-related information includes an image captured by the capture module to control the image acquisition device of the vehicle occupying the parking space and the time of the capture; (2) The storage module is used to store vehicle information, which includes preset license plate characters, license plate characters recognized and obtained by the image character recognition module, and the time when the license plate characters are recognized and obtained by the image character recognition module; (3) The image target detection module is used to perform image detection on the continuous images collected by the image acquisition device, and the image detection includes performing target detection on the license plate on the image containing the license plate; (4) The image character recognition module is used to perform image character recognition, including performing character recognition on the license plate detected by the image target detection module to obtain the license plate characters; (5) The management module is used to manage the work of the modules connected thereto, and also includes license plate character comparison, which is to compare the license plate characters obtained by the image character recognition module with the preset license plate characters. If the license plate characters recognized by the character comparison are not matched with the preset license plate characters during multiple consecutive character comparisons, the image acquisition device is triggered to capture an image of the vehicle occupying the parking space to form vehicle-related information, including the captured image and the time of the capture. The vehicle-related information is stored in the storage module, and the vehicle-related information is sent to the user terminal through the communication module. (6) The capture module is used to control the image acquisition device to capture the vehicle occupying the parking space; (7) The text-to-speech module is used to perform text-to-speech operations, including calling the control module according to the result of the license plate character comparison by the management module, and controlling the voice device to issue a prompt message; The user terminal includes a user terminal device and is configured with a short message service and an email service for receiving vehicle-related information, including data and images, sent from the communication module; The specific steps of the management method of dedicated parking space management by the dedicated parking space autonomous management device based on target detection are as follows: Step 1: The image acquisition device continuously acquires images of the parking space and transmits the images to the image target detection module for target detection of the license plate: (1) If the detection result is valid, that is, the license plate is detected, go to step 2; (2) If the detection result is invalid, that is, the license plate is not detected, go to step 1; Step 2: The image character recognition module performs character recognition on the license plate for the valid detection results. If the character recognition result is valid, that is, the number of recognized valid license plate characters is greater than or equal to the preset number of valid license plate characters, then proceed to step 3; If the character recognition result is valid, that is, the number of recognized valid license plate characters is less than the preset number of valid license plate characters, then go to step 1; Step 3: The management module compares the license plate characters obtained by the image character recognition module with the preset license plate characters: (1) If the characters on the license plate match the preset characters, the text-to-language module calls the control module to control the voice device to emit a corresponding welcome language, and then proceeds to step 4; (2) If the characters do not match the preset license plate characters, the text-to-language module calls the control module to control the voice device to issue a corresponding warning language, and then proceeds to step 5; Step 4: The management module continuously tracks the last recognition time of the license plate image character recognition. When the license plate cannot be recognized again after a preset first time length, the text-to-language module calls the control module to control the voice device to emit a corresponding farewell language, and then enters step 1; Step 5: The management module and the management storage module respectively continuously track the first and last recognition times of the image character recognition of the license plate, and proceed to step 5.1; Step 5.1: When the last recognition time minus the first recognition time exceeds the preset second time length, the text-to-speech module calls the control module to control the voice device to issue a corresponding first warning language, and when the last recognition time minus the first recognition time is between the second time length and the preset third time length: (1) If the vehicle leaves, the management module continuously tracks the last recognition time of the image character recognition of the license plate. When the license plate cannot be recognized again after a first time period, the text-to-speech module calls the control module to control the voice device to emit a corresponding thank you language, and then enters step 1; (2) If the vehicle has not left, that is, the last recognition time minus the first recognition time exceeds the third time length, proceed to step 5.2; Step 5.2: When the last recognition time minus the first recognition time exceeds the third time length, the text-to-speech module calls the control module to control the voice device to issue a corresponding second warning language, and when the last recognition time minus the first recognition time is between the third time length and the preset fourth time length: (1) If the vehicle leaves, the management module and the management storage module continue to track the last recognition time of the image character recognition of the license plate. When the license plate cannot be recognized again after a first time length, the text-to-language module calls the control module to control the voice device to emit a corresponding thank you cooperation language, and then enters step 1; (2) If the vehicle has not left, that is, the last recognition time minus the first recognition time exceeds the fourth time length, proceed to step 5.3; Step 5.3: When the last recognition time minus the first recognition time exceeds the fourth time length, the text-to-speech module calls the control module to control the voice device to issue a corresponding third or final warning language when the last recognition time minus the first recognition time is between the fourth time length and the preset fifth time length: (1) If the vehicle leaves, the management module and the management storage module continue to track the last recognition time of the image character recognition of the license plate. When the license plate cannot be recognized again after a first time length, the text-to-language module calls the control module to control the voice device to emit a corresponding thank you cooperation language, and then enters step 1; (2) If the vehicle has not left, that is, the last recognition time minus the first recognition time exceeds the fifth time length, proceed to step 6; Step 6, the management module manages the capture module, calls the image acquisition device to capture images of illegally parked vehicles, and manages the text-to-speech module, calls the control module, controls the voice device to issue a corresponding illegal parking notification language, manages the communication module, and notifies the user terminal of the captured vehicle images and information via SMS and email, and manages the storage module to store illegal parking images and illegal parking data.
2. The dedicated parking space autonomous management device based on target detection according to claim 1, It is characterized in that The image target detection module adopts the YOLOv5 deep learning algorithm. The YOLOv5 algorithm extracts features from the input image through a convolutional neural network, and uses an anchor frame mechanism to predict the position, size and category of the license plate to achieve target detection of the license plate.
Citation Information
Patent Citations
An automated vehicle parking management system
AU2013293034A1
Occupied parking place supervision system
CN105303879A
Parking stall detection method and device
CN106652551A
GPS and GIS-based method for monitoring waiting vehicles in warehouse regions
CN107886261A
Vehicle behavior information acquiring method, device and terminal
CN108230680A
