Parking lot real-time monitoring system and method based on sensor

Abstract

The invention relates to the technical field of parking lot monitoring, and discloses a parking lot real-time monitoring system and method based on a sensor, and the system comprises a sensor array, a sensor module of an image information collection unit, and a weather compensation module which is used for compensating the influence of weather conditions on the detection precision of the sensor. Comprising an image priority unit, an image processing unit and a parking space state unit, the scheduling module comprises a parking space sorting unit, a parking space scheduling unit and a parking space scheduling feedback unit, and the optimization module comprises a parking space change monitoring model and a parking space prediction unit. And the accuracy of the identification result is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of parking lot monitoring, and discloses a sensor-based real-time parking lot monitoring system and method. Background Art

[0002] The sensor-based real-time parking monitoring system has developed rapidly in recent years. It uses ultrasonic, infrared, laser and other sensor technologies to detect the occupancy of parking spaces. It combines sensors with Internet of Things technology to achieve real-time data transmission and remote monitoring. Users can view the status of parking spaces in real time through mobile phone applications. By analyzing parking data, the system can predict parking needs, optimize the allocation of parking resources, and improve the efficiency of parking lot use. The system can automatically record vehicle entry and exit times, generate statistical reports, and help managers make decisions. However, there are still some shortcomings in the development. For example, there are many different sensors and monitoring systems on the market, and there is a lack of unified technical standards, resulting in poor compatibility between systems, which affects the overall promotion and application. In severe weather conditions (such as rain, snow, haze, etc.), the performance of sensors may be affected, resulting in a decrease in monitoring effect. The low identification of irregular parking in parking lots and the lack of parking allocation result in the inability to feedback parking spaces in parking lots to car owners, resulting in low parking lot management efficiency.

[0003] For example, the existing Chinese patent application with publication number CN106504583A discloses a cloud monitoring method, cloud server and cloud monitoring system for parking space status in a parking lot, wherein the method includes: the parking lot monitoring system collects static parking space status images and uploads them to the cloud server; the cloud server compares the static parking space status images with the existing parking space status models stored in the cloud server, determines and stores the parking space status information; if they are consistent, the parking space status is directly determined, if not, it is manually determined and used as a new parking space status model; the driver reserves an idle parking space through the driver client and sends it to the cloud server, and the parking space status display terminal is updated in real time. This invention can realize information exchange between the parking lot and the driver, and improve the refined management level of the parking lot.

[0004] The above patent only discloses the system framework, but does not indicate how to process the monitoring data. Since all data needs to be transmitted to the cloud via the Internet and then returned, the quality of the network directly affects the response speed and reliability of the system. The deployment and maintenance of such a system requires a high initial investment cost, which may not be affordable for some small parking lots. Summary of the invention

[0005] In order to solve the above technical problems, the main purpose of the present invention is to provide a sensor-based parking lot real-time monitoring system, comprising:

[0006] A sensor module, including a sensor array and an image information acquisition unit;

[0007] A weather compensation module, used to compensate for the influence of weather conditions on the detection accuracy of the sensor, the weather compensation module includes a meteorological data unit, a data correction unit and an environment adaptation unit;

[0008] An image processing module, comprising an image priority unit, an image processing unit, and a parking space status unit, wherein the image priority unit is used to classify images according to their importance and process the images according to the classification results, and the parking space status unit is used to record the real-time status of the parking space;

[0009] A scheduling module, including a parking space sorting unit, a parking space scheduling unit, and a parking space scheduling feedback unit, wherein the parking space scheduling unit is used to schedule parking space information;

[0010] The optimization module includes a parking space change monitoring model and a parking space prediction unit, wherein the parking space change monitoring model is used to predict the change trend of the parking space status;

[0011] The multimodal data module includes an auxiliary unit for assisting in determining the parking space status.

[0012] As a preferred solution of a sensor-based parking lot real-time monitoring system of the present invention, wherein:

[0013] The sensor array includes ultrasonic sensors, infrared sensors, magnetic induction sensors and image sensors for collecting vehicle images and detecting the presence or movement of vehicles;

[0014] If the vehicle exists, both the parking image and the parking space image are collected; if the vehicle does not exist, only the parking space image is collected;

[0015] The parking image includes a vehicle model image and a license plate image;

[0016] The parking space image includes a parking space frame image and a parking space number image.

[0017] As a preferred solution of a sensor-based parking lot real-time monitoring system of the present invention, wherein:

[0018] The data correction unit includes temperature compensation, humidity compensation and precipitation compensation;

[0019] The temperature compensation adjusts the standard temperature of the sensor according to the current air temperature;

[0020] The temperature compensation calculation expression is as follows:

[0021] ;

[0022] in, It is the standard temperature for normal operation of the sensor. is the optimal threshold for normal operation of the sensor, e is the exponential constant, k is the temperature coefficient, T is the actual temperature, is the temperature threshold under the actual temperature T;

[0023] The humidity compensation calculation expression is as follows:

[0024] ;

[0025] in, is the actual transmission power of the sensor, is the standard transmission power of the sensor, is the transmission power coefficient, is the actual humidity, is the standard humidity, is the sensor sensitivity adjustment factor;

[0026] The precipitation compensation is used to provide a compensation factor to adjust the parking space status image processing parameters. The parking space image processing parameters include exposure compensation. The exposure time and exposure gain are processed through exposure compensation. The calculation expression is as follows:

[0027] ;

[0028] ;

[0029] Where F is the exposure time, For the optimal exposure time, is the exposure time coefficient, For the environment affecting the visibility intensity, is the exposure gain, For optimal exposure gain, is the exposure gain factor.

[0030] As a preferred solution of a sensor-based parking lot real-time monitoring system of the present invention, wherein:

[0031] The image priority unit is used to obtain an image priority coefficient, classify the image according to the importance of the image, and process the image according to the classification result;

[0032] The image priority coefficient calculation expression is as follows:

[0033] ;

[0034] in, is the average brightness of image X, is the average brightness of image Y, is the covariance of image X and image Y, which is used to capture the structural information between images. is the variance of image X, is the variance of image Y, is the image priority coefficient, is a constant used to prevent the numerator from being zero, is a constant used to prevent the denominator from being zero.

[0035] As a preferred solution of a sensor-based parking lot real-time monitoring system of the present invention, wherein:

[0036] Setting a priority judgment function for detecting whether the image similarity between a current timestamp and a previous timestamp is lower than a threshold value of 0, wherein the timestamp represents a time interval, for example, the time difference between adjacent times in video processing;

[0037] If the image priority coefficient is less than the threshold value 0, the priority judgment function returns 1, indicating that the difference between the current timestamp and the previous timestamp is large enough. If the image priority coefficient is greater than or equal to the threshold value 0, the priority judgment function returns 0, indicating that the difference between the current timestamp and the previous timestamp does not need to trigger any action and alarm.

[0038] As a preferred solution of a sensor-based parking lot real-time monitoring system of the present invention, wherein:

[0039] The image processing unit is used for image processing, identifying whether the parking space is occupied and whether the parking is in compliance with regulations, and for parking license plate recognition;

[0040] The image processing method comprises:

[0041] S101. Extract feature vectors from images using computer vision;

[0042] S102, training a classifier function using a labeled data set, where the data set includes sample images of regular parking and irregular parking;

[0043] S103, setting a scoring threshold on the output of the classifier function;

[0044] S104, the decision function determines whether parking is in compliance with regulations based on the scoring result of the feature vector by the classifier.

[0045] When the scoring result exceeds the scoring threshold, it is considered that the parking is not in compliance with regulations; otherwise, it is considered that the parking is in compliance with regulations;

[0046] S105, the system needs to continuously process new image frames and make decisions for each image;

[0047] S106. Allow the user or administrator to adjust the scoring threshold or improve the parking space change monitoring model according to actual conditions, and continuously collect data to optimize the model.

[0048] As a preferred solution of a sensor-based parking lot real-time monitoring system of the present invention, wherein:

[0049] The parking space change monitoring model input vector calculation expression is as follows:

[0050] ;

[0051] in, Real-time parking data. For parking space status, is the weather compensation factor;

[0052] The calculation expression of the parking space change monitoring model is as follows:

[0053] ;

[0054] in, is the parking space status at time t+1, is the weight matrix, tanh(·) is the hyperbolic tangent activation function, σ[·] is the sigmoid activation function, b is the bias term, i is the parking space number, M is the total number of parking spaces in the parking lot, and x i is the parking space status of the i-th parking space.

[0055] As a preferred solution of a sensor-based parking lot real-time monitoring system of the present invention, wherein:

[0056] The loss function is used to quantify the gap between the predicted parking space value and the actual parking space value.

[0057] ;

[0058] Among them, t is the time series, log is the mathematical logarithmic function, L is used to output the entropy loss, and J is the parking space prediction value.

[0059] A sensor-based parking lot real-time monitoring method, comprising:

[0060] S1. Collecting parking images and parking space images through sensors, and setting priorities of the parking images and parking space images collected by the sensors through image processing;

[0061] S2. Analyze whether the parking space is vacant and whether illegal parking occurs through the processed parking image and parking space image;

[0062] S3. If the parking space is vacant, parking scheduling is performed directly on the vacant space through the parking space ranking. If the parking space is not vacant, the parking space change trend is predicted through the parking space change monitoring model.

[0063] S4, the parking space change monitoring model outputs the parking space change trend, and inputs the parking space change trend into the parking space prediction unit to predict the parking space availability;

[0064] S5. Feedback the parking space situation to the car owner and provide the car owner with parking space information.

[0065] A preferred solution of a sensor-based real-time parking lot monitoring method, wherein:

[0066] The parking space prediction unit is used to receive and output the parking space prediction value output by the parking space change monitoring model after loss compensation.

[0067] Beneficial effects of the present invention:

[0068] The present invention discloses a real-time parking lot monitoring system that integrates image priority processing technology and a neural network model. The system collects data through sensors to identify the status of parking spaces in the parking lot, processes image edge information, and establishes image priority to simplify data calculation. The response speed and reliability of the system are improved, and the cost of deployment and maintenance is reduced. The system helps car owners to schedule parking spaces and makes it convenient for car owners to find parking spaces in the first place. The neural network model is used to optimize the recognition of parking space status and ensure the accuracy of recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0070] Figure 1 This is a system composition diagram of a sensor-based parking lot real-time monitoring system of the present invention;

[0071] Figure 2 This is a flow chart of a sensor-based real-time parking lot monitoring method of the present invention;

[0072] Figure 3 The present invention is a flowchart of an image processing method. DETAILED DESCRIPTION

[0073] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0074] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0075] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0076] Example 1

[0077] like Figure 1 As shown, a sensor-based parking lot real-time monitoring system includes:

[0078] A sensor module, including a sensor array and an image information acquisition unit;

[0079] The sensor array includes, but is not limited to, ultrasonic sensors, infrared sensors, magnetic induction sensors, and image sensors, etc., for collecting vehicle images and detecting the presence or movement of the vehicle;

[0080] Ultrasonic sensors are used to measure the distance between the vehicle and the edge of the parking space to help determine whether the vehicle is parked in the space.

[0081] Infrared sensors are used to detect whether an object (such as a vehicle) enters or leaves the parking area.

[0082] The magnetic induction sensor is installed under the ground of the parking space. When a metal object (such as a car) approaches, the magnetic field changes, thereby detecting the presence of the vehicle.

[0083] The image sensor is integrated in the image information acquisition unit and is used to capture images of parking spaces and vehicles.

[0084] When any sensor detects that a vehicle enters or leaves a parking space, the image information acquisition unit is triggered to capture an image;

[0085] If the vehicle exists, both the parking image and the parking space image are collected; if the vehicle does not exist, only the parking space image is collected;

[0086] Parking images include vehicle model images and license plate images;

[0087] The parking space image includes a parking space border image and a parking space number image;

[0088] Design considerations include: ensuring that the location of sensors and cameras covers the entire parking space, especially the key parts of the vehicle, selecting cameras with sufficient resolution and ensuring clear images under various lighting conditions, designing a reliable data transmission mechanism to ensure that image data can be efficiently transmitted between the sensor and the image processing module, and considering energy consumption issues, selecting sensors and cameras with energy-saving modes to enter low-power modes when there is no vehicle activity;

[0089] The weather compensation module is used to compensate for the impact of weather conditions on the sensor detection accuracy.

[0090] The weather compensation module includes a meteorological data unit, a data correction unit and an environmental adaptation unit;

[0091] The meteorological data unit is used to access real-time weather data, such as temperature, humidity, wind speed, rainfall, etc.

[0092] The data correction unit adjusts the detection threshold and algorithm parameters of the sensor based on weather data to adapt to different weather conditions;

[0093] Further, the data correction unit includes temperature compensation, humidity compensation and precipitation compensation;

[0094] Temperature compensation adjusts the sensor's standard temperature according to the current air temperature;

[0095] The temperature compensation calculation expression is as follows:

[0096] ;

[0097] in, It is the standard temperature for the sensor to work normally. is the optimal threshold for normal operation of the sensor, e is the exponential constant, k is the temperature coefficient, T is the actual temperature, is the temperature threshold under the actual temperature T;

[0098] Humidity compensation: If the humidity is high, the reflected signal of the ultrasonic sensor is weakened. The humidity compensation algorithm is used to adjust the transmit power and receive sensitivity of the ultrasonic sensor according to the humidity level;

[0099] The humidity compensation calculation expression is as follows:

[0100] ;

[0101] in, is the actual transmission power of the sensor, is the standard transmission power of the sensor, is the transmission power coefficient, is the actual humidity, is the standard humidity, is the sensor sensitivity adjustment factor;

[0102] Rainy and snowy weather will affect the imaging effect of the image sensor. Precipitation compensation is used to provide a compensation factor to adjust the parking status image processing parameters to improve image quality. The parking status image processing parameters include exposure compensation, which processes the exposure time and exposure gain;

[0103] ;

[0104] ;

[0105] Where F is the exposure time, For the optimal exposure time, is the exposure time coefficient, For the environment affecting the visibility intensity, is the exposure gain, For optimal exposure gain, is the exposure gain factor;

[0106] The environmental adaptation unit formulates corresponding compensation strategies according to different weather conditions to ensure the consistency and reliability of sensor detection;

[0107] By combining different weather compensation factors, a weather compensation factor is output;

[0108] Through precise mathematical models and compensation mechanisms, high measurement accuracy can be maintained in different environments, and it can quickly respond and adapt to changing environmental conditions, including extreme weather. Users do not need to manually adjust settings, and the system will automatically optimize according to current environmental conditions. Appropriate compensation measures can reduce equipment wear caused by environmental stress and help extend its service life. Regardless of how external conditions change, it can always provide stable and reliable data, enhancing the practicality and competitiveness of the real-time parking lot monitoring system.

[0109] By using different mathematical models to design temperature compensation, humidity compensation, and precipitation compensation algorithms, the impact of different environmental factors on sensor performance can be more accurately reflected, and these models can be further adjusted and optimized according to actual application scenarios;

[0110] The auxiliary unit uses various types of data (such as sound, temperature, etc.) to assist in determining the parking space status;

[0111] Auxiliary units include sound sensors, temperature sensors, and other sensors;

[0112] The sound sensor is used to detect the sounds in the parking lot, such as the sound of door opening and closing, the sound of engine starting, etc.

[0113] The temperature sensor is used to detect the temperature changes around the parking space, which indirectly reflects whether there is a vehicle;

[0114] Other sensors, such as humidity sensors and wind speed sensors, can also provide auxiliary information;

[0115] An image processing module, including an image priority unit, an image processing unit, and a parking space status unit;

[0116] Among them, the image priority unit is used to obtain the image priority coefficient, classify the images according to their importance, and give priority to key images;

[0117] Covariance is used to capture the structural similarity or correlation between image X and image Y. If two images are very different in structure, it means that they each contain unique information, which may be one of the key factors for priority processing. It represents the variance of image X and image Y, respectively, which is used to measure the degree of variation of pixel values ​​within the image. High variance usually means that the image has more details or changes, and therefore may be more important. In order to ensure the stability of the formula, especially when the numerator or denominator is close to zero, a small constant is added to prevent division by zero errors and maintain the rationality of the calculation results. The final image priority coefficient is the result of comprehensively considering factors such as brightness, structural similarity, and internal changes. The image priority coefficient helps the system decide which images are critical and should be prioritized.

[0118] The image priority coefficient calculation expression is as follows:

[0119] ;

[0120] in, is the average brightness of image X, is the average brightness of image Y, is the covariance of image X and image Y, which is used to capture the structural information between images. is the variance of image X, is the variance of image Y, is the image priority coefficient, is a constant used to prevent the numerator from being zero, is a constant used to prevent the denominator from being 0;

[0121] By automatically evaluating the importance of images, the system can optimize resource usage and focus limited processing power on the most important tasks. Prioritizing critical images allows the system to respond to emergencies more quickly or provide more valuable information, thereby improving overall work efficiency. Through in-depth analysis of image content, the system can better understand the actual meaning of each image and help make more accurate decisions. The algorithm can adjust parameters according to different application scenarios and is suitable for various types of image processing tasks.

[0122] Furthermore, the calculation expression of the average brightness μ is as follows:

[0123] ;

[0124] in, is the height of the image, is the width of the image, is the horizontal coordinate of the image, D is the vertical coordinate of the image; I (L, D) represents the overall brightness of the image;

[0125] variance The calculation expression is as follows:

[0126] ;

[0127] in, is the average brightness of image X, is the average brightness of image Y;

[0128] Furthermore, and Both represent the average brightness of image X;

[0129] and Both represent the average brightness of image Y;

[0130] Set a priority judgment function to detect whether the image similarity between the current timestamp and the previous timestamp is lower than the threshold 0. The timestamp here represents the time interval, for example, in video processing it may be the time difference between adjacent times;

[0131] By comparing image similarities, frames with little change can be skipped, thus saving computing resources. When a significant change is detected (such as motion), the processing priority of the frame can be immediately increased to ensure that important information is processed in a timely manner. In network transmission scenarios, reducing unnecessary data transmission can effectively reduce bandwidth consumption, especially in monitoring and other real-time applications. This method can help the system respond to emergencies more quickly. By automatically identifying which frames are more important, the system can more intelligently allocate limited processing resources to ensure that key frames receive adequate attention.

[0132] If the image priority coefficient is less than the threshold value 0, the priority judgment function returns 1, which means that the difference between the current timestamp and the previous timestamp is large enough to require attention or some action; if the image priority coefficient is greater than or equal to the threshold value 0, the priority judgment function returns 0, indicating that the difference between the current timestamp and the previous timestamp does not need to trigger any action or alarm;

[0133] The image processing unit is used to process images, identify whether the parking space is occupied and whether the parking is in compliance with regulations, and recognize parking license plates;

[0134] like Figure 3 As shown, further, the image processing method includes:

[0135] S101. Use computer vision technology to extract useful feature vectors from the image. Features include but are not limited to the outline of the vehicle, the position of the vehicle relative to the parking space, the posture of the vehicle, etc.;

[0136] The images include parking images and parking space images;

[0137] Among them, the parking area is first separated from the background by image segmentation, and then useful features such as color, texture, shape, etc. are extracted from the segmented parking area by feature extraction;

[0138] S102, training a classifier function using a labeled data set, where the data set should contain sample images of standardized parking and irregular parking;

[0139] Furthermore, the classifier may be a support vector machine (SVM), a decision tree, a random forest, a logistic regression or other machine learning model, or a deep learning model, and the training process should adjust the model parameters so that it can distinguish between regular parking and irregular parking situations;

[0140] Use classifiers (such as support vector machines (SVM), random forests (RF), convolutional neural networks (CNN), etc.) to determine whether a parking space is occupied;

[0141] S103, setting a scoring threshold on the output of the classifier function;

[0142] Among them, the scoring threshold can be adjusted according to actual conditions and needs, and the selection of the scoring threshold may depend on the performance of the model during cross-validation to ensure that the model is neither too sensitive nor too slow;

[0143] S104. The decision function determines whether parking is in compliance with regulations based on the scoring result of the feature vector by the classifier.

[0144] When the scoring result exceeds the scoring threshold, it is considered that the parking is not in compliance with regulations; otherwise, it is considered that the parking is in compliance with regulations;

[0145] Further, determining whether the vehicle is correctly parked in the parking space includes contour detection and position comparison;

[0146] Furthermore, contour detection uses edge detection algorithms (such as Canny edge detection) to detect vehicle contours;

[0147] Position comparison compares the detected vehicle outline with the parking space border to determine whether the vehicle is completely inside the parking space;

[0148] S105. For real-time applications, the system needs to continuously process new image frames and make decisions quickly;

[0149] S106. Feedback mechanism, which allows users or administrators to adjust the scoring threshold or improve the parking space change monitoring model according to actual conditions. By continuously collecting data and optimizing the model, the system's ability to judge whether parking is standardized can be continuously improved;

[0150] Parking license plate recognition includes license plate location, character segmentation and character recognition;

[0151] Furthermore, license plate location uses a license plate location algorithm (such as Haar feature cascade classifier) ​​to locate the license plate position;

[0152] Character segmentation divides the license plate area into individual characters;

[0153] Character recognition: Use OCR technology (such as Tesseract OCR) to recognize each character;

[0154] Parking license plate recognition technology uses existing technology to meet image acquisition requirements;

[0155] The parking space status unit is used to record the real-time status of the parking space;

[0156] The scheduling module includes a parking space sorting unit, a parking space scheduling unit, and a parking space scheduling feedback unit;

[0157] The parking space sorting unit is used to sort the parking spaces according to the parking space status and reservation information;

[0158] If the parking space is not occupied and not reserved, it has the highest priority;

[0159] If the parking space is not occupied but has been reserved, it has medium priority;

[0160] In other cases, the priority is the lowest;

[0161] The priority is expressed by the following calculation expression:

[0162] ;

[0163] in, For parking space status, For the reservation status, For priority.

[0164] and They are all parameters representing the same parking space status;

[0165] The parking space scheduling unit is used to arrange the parking of vehicles according to the sorting result, including obtaining the sorted parking space list from the parking space sorting unit, traversing the sorted parking space list, finding the first free and unreserved parking space, allocating it to the vehicle that needs to park, and sending the allocated parking space information to the owner to indicate the specific parking location;

[0166] The parking space scheduling unit is used to schedule parking space information and output it to the car owner who needs to park;

[0167] The parking space scheduling feedback unit includes: receiving confirmation information from the car owner and updating the status of the parking space based on the confirmation situation, receiving confirmation information sent by the car owner after arriving at the parking space through the car owner's terminal application, and updating the parking space status record in the parking space sorting unit based on the car owner's confirmation information. If the car owner fails to arrive at the parking space on time, the system needs to reallocate the parking space or prompt the car owner to make a new reservation.

[0168] Optimization module, including parking space change monitoring model and parking space prediction unit;

[0169] The parking space change monitoring model is used to predict the changing trend of parking space status;

[0170] The parking space change monitoring model input vector calculation expression is as follows:

[0171] ;

[0172] in, Real-time parking data. For parking space status, is the weather compensation factor;

[0173] By using real-time parking space data, parking space status and weather compensation factors, the model can comprehensively consider the impact of multiple factors on parking space status and improve the accuracy of prediction. By introducing weather compensation factors, the model can better adapt to changes in parking demand under different weather conditions, thereby improving prediction accuracy.

[0174] The calculation expression of the parking space change monitoring model is as follows:

[0175] ;

[0176] in, is the parking space status at time t+1, is the weight matrix, tanh(·) is the hyperbolic tangent activation function, σ[·] is the sigmoid activation function, b is the bias term, i is the parking space number, M is the total number of parking spaces in the parking lot, and x i is the parking space status of the i-th parking space;

[0177] Furthermore, the parking space change monitoring model includes multiple gate units, each of which is responsible for processing a time step in the time series. The gate unit controls the flow of information through a gating mechanism;

[0178] The gating mechanism includes controlling the input gate, forget gate, output gate, candidate unit, state unit, loss function and hidden state;

[0179] The input gate is used to control the impact of the current input on the unit state;

[0180] The forget gate is used to decide how much of the previous cell state to retain;

[0181] The output gate is used to determine the impact of the current cell state on the output:

[0182] Candidate cells are used to update cell status:

[0183] The loss function is used to quantify the gap between the predicted value and the true value;

[0184] ;

[0185] Among them, t is the time series, log is the mathematical logarithmic function, L is used to output the entropy loss, and J is the parking space prediction value;

[0186] These gating mechanisms allow the model to selectively retain or ignore information and effectively handle long-term dependency issues. For example, the input gate controls the impact of the current input on the cell state; the forget gate determines how much of the previous cell state is retained; and the output gate determines the impact of the current cell state on the output. This helps the model focus on truly important information and reduce noise interference. Candidate cells are used to update cell states, ensuring that the model can continue to learn and adjust its internal representation to adapt to new data patterns. The entropy loss function L is used to quantify the gap between the predicted value and the true value, guiding the model to optimize parameters and make predictions more accurate.

[0187] The parking space prediction unit provides a prediction of the future availability of parking spaces;

[0188] The parking space prediction unit is used to receive the prediction value output by the parking space change monitoring model after loss compensation, and output it through the parking space prediction unit;

[0189] The parking space change monitoring model is a mathematical representation of the logic of the result output by the gate control mechanism. The algorithm structure can be changed according to the actual production application.

[0190] Redundancy module, used to set up network redundancy scheme and backup server;

[0191] Adopt dual-channel network connection (wired + wireless) and backup server to ensure that the system can still operate normally when the network fails;

[0192] For example, use multiple network connections (wired + wireless), backup servers, etc. to ensure that the system can maintain normal operation even in poor network conditions;

[0193] Security module, used to strengthen system security design and protect user data security;

[0194] Data encryption transmission technology, access control mechanisms, firewalls and other means are used to ensure the security and integrity of data during transmission.

[0195] Furthermore, the parking space prediction value includes whether the parking space will be vacant, occupied, or under maintenance in the future;

[0196] Parking images include but are not limited to whether parking is in compliance with regulations;

[0197] The actual value of the parking space is whether the parking space is vacant or occupied, or is under maintenance at the current timestamp.

[0198] Example 2

[0199] like Figure 2 As shown, a sensor-based parking lot real-time monitoring method implementation process includes:

[0200] S1. Collecting parking images and parking space images through sensors, and setting priorities of the parking images and parking space images collected by the sensors through image processing;

[0201] S2. Analyze whether the parking space is vacant and whether illegal parking occurs through the processed parking image and parking space image;

[0202] S3. If the parking space is vacant, parking scheduling is performed directly on the vacant space through the parking space ranking. If the parking space is not vacant, the parking space change trend is predicted through the parking space change monitoring model.

[0203] S4, the parking space change monitoring model outputs the parking space change trend, and inputs the parking space change trend into the parking space prediction unit to predict the parking space availability;

[0204] S5. Feedback the parking space situation to the car owner and provide the car owner with parking space information.

[0205] Importantly, it should be noted that the construction and arrangement of the present application shown in a plurality of different exemplary embodiments are only exemplary. Although only two embodiments are described in detail in this disclosure, it should be readily understood by those who refer to this disclosure that many modifications are possible, for example, the size, scale, structure, shape and ratio of various elements, and parameter values ​​(e.g., temperature, pressure, etc.), installation arrangement, use of materials, color, directional changes, etc., without substantially departing from the novel teachings and advantages of the subject matter described in the application. For example, the element shown as integrally formed can be composed of multiple parts or elements, the position of the element can be inverted or otherwise changed, and the nature or number or position of the discrete element can be changed or changed. Therefore, all such modifications are intended to be included in the scope of the present invention. The order or sequence of any process or method steps can be changed or reordered according to alternative embodiments. Any "device plus function" clause is intended to cover the structure of the execution function described in this article, and is not only structurally equivalent but also equivalent structure. Without departing from the scope of the present invention, other replacements, modifications, changes and omissions can be made in the design, operating conditions and arrangement of the exemplary embodiment. Therefore, the invention is not limited to a specific embodiment, but extends to numerous modifications still falling within the scope of the appended claims.

[0206] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment may not be described (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those features that are not relevant to implementing the invention).

[0207] It should be understood that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but for those of ordinary skill having the benefit of this disclosure, the development effort will be a routine task of design, fabrication, and production without undue experimentation.

[0208] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A sensor-based parking lot real-time monitoring system, characterized in that: include: A sensor module, including a sensor array and an image information acquisition unit; A weather compensation module, used to compensate for the influence of weather conditions on the detection accuracy of the sensor, the weather compensation module includes a meteorological data unit, a data correction unit and an environment adaptation unit; An image processing module, comprising an image priority unit, an image processing unit, and a parking space status unit, wherein the image priority unit is used to classify images according to their importance and process the images according to the classification results, and the parking space status unit is used to record the real-time status of the parking space; A scheduling module, including a parking space sorting unit, a parking space scheduling unit, and a parking space scheduling feedback unit, wherein the parking space scheduling unit is used to schedule parking space information; The optimization module includes a parking space change monitoring model and a parking space prediction unit, wherein the parking space change monitoring model is used to predict the change trend of the parking space status; The multimodal data module includes an auxiliary unit for assisting in determining the parking space status.

2. A sensor-based parking lot real-time monitoring system according to claim 1, characterized in that: The sensor array includes ultrasonic sensors, infrared sensors, magnetic induction sensors and image sensors for collecting vehicle images and detecting the presence or movement of vehicles; If the vehicle exists, both the parking image and the parking space image are collected; if the vehicle does not exist, only the parking space image is collected; The parking image includes a vehicle model image and a license plate image; The parking space image includes a parking space frame image and a parking space number image.

3. A sensor-based parking lot real-time monitoring system according to claim 2, characterized in that: The data correction unit includes temperature compensation, humidity compensation and precipitation compensation; The temperature compensation adjusts the standard temperature of the sensor according to the current air temperature; The temperature compensation calculation expression is as follows: ； in, It is the standard temperature for normal operation of the sensor. is the optimal threshold for normal operation of the sensor, e is the exponential constant, k is the temperature coefficient, T is the actual temperature, is the temperature threshold under the actual temperature T; The humidity compensation calculation expression is as follows: ； in, is the actual transmission power of the sensor, is the standard transmission power of the sensor, is the transmission power coefficient, is the actual humidity, is the standard humidity, is the sensor sensitivity adjustment factor; The precipitation compensation is used to provide a compensation factor to adjust the parking space status image processing parameters. The parking space status image processing parameters include exposure compensation. The exposure time and exposure gain are processed through exposure compensation. The calculation expression is as follows: ； ； Where F is the exposure time, For the optimal exposure time, is the exposure time coefficient, For the environment affecting the visibility intensity, is the exposure gain, For optimal exposure gain, is the exposure gain factor.

4. A sensor-based parking lot real-time monitoring system according to claim 3, characterized in that: The image priority unit is used to obtain an image priority coefficient, classify the image according to the importance of the image, and process the image according to the classification result; The calculation expression of the image priority coefficient is as follows: ； in, is the average brightness of image X, is the average brightness of image Y, is the covariance of image X and image Y, which is used to capture the structural information between images. is the variance of image X, is the variance of image Y, is the image priority coefficient, is a constant used to prevent the numerator from being zero, is a constant used to prevent the denominator from being zero.

5. A sensor-based parking lot real-time monitoring system according to claim 4, characterized in that: Setting a priority judgment function for detecting whether the image similarity between a current timestamp and a previous timestamp is lower than a threshold value of 0, wherein the timestamp represents a time interval; If the image priority coefficient is less than the threshold value 0, the priority judgment function returns 1, indicating that the difference between the current timestamp and the previous timestamp is large enough. If the image priority coefficient is greater than or equal to the threshold value 0, the priority judgment function returns 0, indicating that the difference between the current timestamp and the previous timestamp does not need to trigger any action and alarm.

6. A sensor-based parking lot real-time monitoring system according to claim 5, characterized in that: The image processing unit is used for image processing, identifying whether the parking space is occupied and whether the parking is in compliance with regulations, and for parking license plate recognition; The image processing method comprises: S101. Extract feature vectors from images using computer vision; S102, training a classifier function using a labeled data set, where the data set includes sample images of regular parking and irregular parking; S103, setting a scoring threshold on the output of the classifier function; S104, the decision function determines whether parking is in compliance with regulations based on the scoring result of the feature vector by the classifier. When the scoring result exceeds the scoring threshold, it is considered that the parking is not in compliance with regulations; otherwise, it is considered that the parking is in compliance with regulations; S105, the system needs to continuously process new image frames and make decisions for each image; S106. Allow the user or administrator to adjust the scoring threshold or improve the parking space change monitoring model according to actual conditions, and continuously collect data and optimize the model.

7. A sensor-based parking lot real-time monitoring system according to claim 6, characterized in that: The parking space change monitoring model input vector calculation expression is as follows: ； in, Real-time parking space data. For parking space status, is the weather compensation factor; The calculation expression of the parking space change monitoring model is as follows: ； in, is the parking space status at time t+1, is the weight matrix, tanh(·) is the hyperbolic tangent activation function, σ[·] is the sigmoid activation function, b is the bias term, i is the parking space number, M is the total number of parking spaces in the parking lot, and x i is the parking space status of the i-th parking space.

8. A sensor-based parking lot real-time monitoring system according to claim 7, characterized in that: The loss function is used to quantify the gap between the predicted parking space value and the actual parking space value. ； Among them, t is the time series, log is the mathematical logarithmic function, L is used to output the entropy loss, and J is the parking space prediction value.

9. A sensor-based parking lot real-time monitoring method, implemented based on a sensor-based parking lot real-time monitoring system according to any one of claims 1 to 8, characterized in that: include: S1. Collecting parking images and parking space images through sensors, and setting priorities of the parking images and parking space images collected by the sensors through image processing; S2. Analyze whether the parking space is vacant and whether illegal parking occurs through the processed parking image and parking space image; S3. If the parking space is vacant, parking scheduling is performed directly on the vacant space through the parking space ranking. If the parking space is not vacant, the parking space change trend is predicted through the parking space change monitoring model. S4, the parking space change monitoring model outputs the parking space change trend, and inputs the parking space change trend into the parking space prediction unit to predict the parking space availability; S5. Feedback the parking space situation to the car owner and provide the car owner with parking space information.

10. A sensor-based real-time parking lot monitoring method according to claim 9, characterized in that: The parking space prediction unit is used to receive and output the parking space prediction value output by the parking space change monitoring model after loss compensation.

Images