Parking lot parking space occupation monitoring and abnormity early warning system integrating geomagnetism and vision

Abstract

The invention discloses a parking lot parking space occupation monitoring and abnormity early warning system based on geomagnetism and vision fusion. The system comprises a data preprocessing unit, a data fusion unit, a state determination unit and an abnormity early warning unit. The data preprocessing unit is used for acquiring geomagnetic original data and visual original data of a target parking space of a parking lot and preprocessing the geomagnetic original data and the visual original data respectively; the invention relates to the technical field of parking lot parking space monitoring. According to the parking lot parking space occupancy monitoring and abnormity early warning system based on geomagnetism and vision fusion, accurate monitoring of parking lot parking space occupancy states is realized through data fusion of the geomagnetism and vision sensors, the system fuses geomagnetism data and vision data through a Kalman filtering algorithm, and the parking space occupancy monitoring and abnormity early warning system based on geomagnetism and vision fusion is realized on the basis of guaranteeing the monitoring precision. The fusion feature value of the parking space state is provided, and the judgment accuracy of the parking space occupancy state is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of parking lot space monitoring, more particularly, to a parking lot space occupation monitoring and abnormal early warning system based on fusion of geomagnetic and visual data. BACKGROUND

[0002] With the acceleration of urbanization and the growth of car ownership, the problem of parking difficulty is increasingly prominent. Parking lot management, as a key component of intelligent transportation systems, is of great significance to improve parking efficiency and reduce traffic congestion. Most existing parking lot space monitoring systems use single technical means, such as geomagnetic sensors or visual sensors. However, these systems have certain limitations in complex environments. For example, geomagnetic sensors are easily disturbed by the environment, resulting in inaccurate data; while pure visual systems may not work reliably in low light or complex weather conditions.

[0003] To solve this problem, researchers have proposed various improvements to parking space monitoring systems, among which the fusion of different sensing technologies, such as geomagnetic and visual data, has gradually become the mainstream. By fusing geomagnetic data and visual data, the advantages of the two different technologies can be utilized to make up for the shortcomings of a single technology, thereby improving the accuracy and robustness of parking space occupation monitoring.

[0004] However, existing geomagnetic and visual fusion systems mostly focus on data fusion and basic monitoring, lacking timely warning and accurate judgment of abnormal parking space status, making it difficult to meet the needs of efficient and intelligent parking management. In the existing technology, there is a lack of early warning mechanism for abnormal parking space status, resulting in monitoring results that can only provide parking occupancy status information and cannot provide real-time warning signals for abnormal behavior. SUMMARY

[0005] The present application aims to provide a parking lot space occupation monitoring and abnormal early warning system based on fusion of geomagnetic and visual data, which solves the problem of lack of early warning mechanism for abnormal parking space status in the prior art, resulting in monitoring results that can only provide parking occupancy status information and cannot provide real-time warning signals for abnormal behavior, failing to meet the use requirements.

[0006] The present application achieves the above-mentioned purpose through the following technical solutions: a parking lot space occupation monitoring and abnormal early warning system based on fusion of geomagnetic and visual data, the system comprising: a data preprocessing unit, a data fusion unit, a state determination unit, and an abnormal early warning unit; The data preprocessing unit is configured to obtain and preprocess the geomagnetic raw data and visual raw data of the target parking space in the parking lot. The data fusion unit is configured to perform fusion calculation on the preprocessed geomagnetic data and visual data through a Kalman filter data fusion algorithm to obtain parking space state fusion feature values. The status determination unit is used to determine the occupancy status of the target parking space based on the comparison between the parking space status fusion feature value and the preset determination threshold. The abnormal warning unit is used to trigger and output an abnormal warning signal when the parking space occupancy status meets the abnormal conditions.

[0007] Furthermore, the raw geomagnetic data is multi-dimensional data collected by a geomagnetic sensor, including at least: The magnetic field strength has x-axis, y-axis, and z-axis components, with a data dimension of K and a data acquisition time step of T. The raw geomagnetic data is stored in the form of a K×T dimension matrix.

[0008] Furthermore, the geomagnetic data preprocessing includes: Normalization is performed to eliminate the dimensional differences between geomagnetic data of different dimensions, resulting in standardized geomagnetic data. The normalization process is calculated based on the mean and standard deviation of geomagnetic data in each dimension.

[0009] Furthermore, The raw visual data is the target parking space image obtained by the image acquisition device deployed above the parking space; Visual data preprocessing includes image grayscale conversion, denoising, region of interest extraction, target feature extraction, and feature parameter standardization to obtain visual feature vectors.

[0010] Furthermore, the target features include at least the target contour area and the average gray level; The area of ​​the target contour is calculated by pixel counting after the contour is extracted by an edge detection algorithm. The mean gray level is the average gray level of all pixels within the region of interest.

[0011] Furthermore, the Kalman filter data fusion algorithm includes: Construct the state equation and the observation equation; Data fusion is achieved through iterative calculations in the prediction and update steps; The state equation includes the parking space occupancy state value and the rate of change of the state; The observation vector of the observation equation is a weighted sum of the mean of geomagnetic data and the visual feature vector.

[0012] Furthermore, the prediction step calculates the predicted state value at time t and the prediction error covariance matrix; The update step involves calculating the Kalman gain, state update, and error covariance update. The parking space occupancy status value is extracted from the updated state vector and used as the parking space state fusion feature value.

[0013] Furthermore, the preset judgment threshold is calculated based on the mean and standard deviation of the fused feature values ​​in the parking space vacancy status in historical monitoring data. By comparing the fused feature values ​​with the preset judgment threshold, the parking space is determined to be occupied or vacant.

[0014] Furthermore, the abnormal conditions include: a first scenario and a second scenario: The first scenario is that within a continuously preset number of time steps, the parking space status conflicts with historical statistical patterns; The second scenario is when the fusion feature value of the parking space status exceeds the preset normal threshold range, and the preset number of time steps is not less than 5.

[0015] Furthermore, in the standardization process of the visual feature vector, the target contour area is normalized by comparing it with the contour area threshold of the maximum target that the preset parking space can accommodate, and the grayscale mean is normalized by comparing it with the maximum grayscale value.

[0016] The beneficial effects of this invention are as follows: 1. This method combines the advantages of both geomagnetic and visual monitoring. Geomagnetic data reflects changes in the geomagnetic field caused by vehicles, while visual data provides intuitive parking space image information. A Kalman filter data fusion algorithm is used to fuse the preprocessed geomagnetic and visual data to obtain a fused feature value for the parking space status. This fused feature value is then compared with a preset judgment threshold to determine the parking space occupancy status. This effectively overcomes the susceptibility to interference inherent in single monitoring methods and significantly improves the accuracy of parking space occupancy status determination.

[0017] 2. Magnetic monitoring and visual monitoring complement each other. When magnetic monitoring is interfered with by surrounding metal objects or visual monitoring is affected by factors such as light and weather, the other monitoring method can still provide effective information. Through data fusion, the system can comprehensively analyze the two types of data, reduce the impact of errors from a single data source on the final judgment result, enhance the system's anti-interference ability under different environmental conditions, and ensure the reliability of parking space monitoring.

[0018] 3. The system is equipped with an anomaly warning unit. When the parking space occupancy status meets the abnormal conditions, such as the parking space status conflicting with historical statistical patterns within a preset number of time steps, or the fused characteristic value of the parking space status exceeding the preset normal threshold range, an anomaly warning signal can be triggered and output in a timely manner. This allows parking lot managers to promptly detect and handle abnormal situations, such as illegal parking or equipment failure, thereby improving the management level and operational efficiency of the parking lot.

[0019] 4. The raw geomagnetic and visual data were preprocessed separately. Geomagnetic data preprocessing included normalization, which eliminated the dimensional differences between geomagnetic data of different dimensions. Visual data preprocessing included steps such as image grayscale conversion, noise reduction, region of interest extraction, target feature extraction, and feature parameter standardization to obtain visual feature vectors. This optimized the data quality and provided an accurate and reliable data foundation for subsequent data fusion and state determination, further improving the system performance.

[0020] 5. The preset judgment threshold is calculated based on the mean and standard deviation of the fused feature values ​​of parking space vacancy status in historical monitoring data. This threshold setting method based on historical data can fully consider the actual use of parking spaces and environmental changes, making the judgment of parking space occupancy status more reasonable and accurate, and reducing the occurrence of misjudgment and missed judgment. Attached Figure Description

[0021] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Fig. 1 This is a system block diagram of the present invention; Fig. 2 This is a detailed flowchart of the data preprocessing unit of the present invention; Fig. 3 This is a flowchart of the Kalman filter fusion and state determination process of the present invention. Detailed Implementation

[0022] The present application will now be described in further detail with reference to the accompanying drawings. It should be noted that the following specific embodiments are only used to further illustrate the present application and should not be construed as limiting the scope of protection of the present application. Those skilled in the art can make some non-essential improvements and adjustments to the present application based on the above application content.

[0023] Example 1: Please see Figs. 1-3 This invention provides a technical solution: a parking lot occupancy monitoring and anomaly early warning system that integrates geomagnetism and vision, the system comprising: Data preprocessing unit, data fusion unit, status determination unit, and anomaly early warning unit; The data preprocessing unit is used to acquire the original geomagnetic data and visual data of the target parking space in the parking lot, and to preprocess them respectively. The data consists of several parts: Geomagnetic raw data, which is collected by geomagnetic sensors and reflects changes in the magnetic field of parking spaces. When a vehicle enters or leaves a parking space, it causes changes in the surrounding magnetic field. The geomagnetic sensors convert these changes into electrical signals, thus forming the raw data. Visual raw data consists of images or videos of parking spaces captured by cameras or other visual devices. This data contains visual information about the parking spaces, such as whether a vehicle is parked there and the vehicle's appearance. Preprocessing involves a series of operations on the raw data to remove noise, improve data quality, and standardize data formats for subsequent data fusion and analysis. For example, geomagnetic raw data may require filtering to eliminate interference signals; visual raw data may require image enhancement, target detection, and other preprocessing steps. The data fusion unit is used to perform fusion calculations on preprocessed geomagnetic data and visual data using the Kalman filter data fusion algorithm to obtain fusion feature values ​​of parking space status. Among them, the Kalman filter data fusion algorithm is a recursive algorithm for estimating the state of dynamic systems. It can combine measurement data from multiple sensors and estimate the system state more accurately in the presence of noise through prediction and update steps. In this system, it fuses information from two different sources, geomagnetic data and visual data, giving full play to the advantages of both technologies and improving the accuracy of parking space status judgment. The parking space status fusion feature value is a comprehensive index obtained after data fusion calculation. It integrates information from both geomagnetic and visual aspects and can more comprehensively and accurately reflect the actual status of the parking space. The status determination unit determines the occupancy status of the target parking space by comparing the fused feature value of the parking space status with a preset determination threshold. Among them, the preset judgment threshold is a boundary value for judging the status of parking spaces that is pre-set based on actual needs and experience. A range of fused feature values ​​is set. When the feature value is within this range, the parking space is judged to be occupied. If it is outside this range, it is judged to be vacant. Occupied status refers to whether the parking space is occupied by a vehicle. It is usually divided into two situations: occupied and vacant. The abnormal warning unit triggers and outputs an abnormal warning signal when the parking space occupancy status meets abnormal conditions; Among them, abnormal conditions are predefined rules or conditions used to determine whether the status of a parking space is abnormal. For example, situations such as a parking space frequently switching between occupied and vacant status in a short period of time, or a parking space being illegally occupied, may be set as abnormal conditions. An abnormal warning signal is a signal used to indicate that the status of a parking space is abnormal. It can take various forms such as sound alarms, flashing lights, and SMS notifications, so that relevant personnel can take timely measures.

[0024] It should be noted that during use, the data preprocessing unit processes the raw geomagnetic and visual data separately, removing noise and improving quality to provide a reliable foundation for subsequent analysis. The data fusion unit uses a Kalman filter algorithm to fuse the two data, combining the advantages of both technologies to overcome the limitations of a single sensor and obtain more accurate and comprehensive fused feature values ​​of parking space status, thus improving judgment accuracy. The status determination unit determines the occupancy status by comparing the fused feature values ​​with preset thresholds, which is simple, efficient, and accurate. The anomaly warning unit promptly triggers signal output when the parking space status meets abnormal conditions, quickly detecting anomalies such as illegal occupancy and frequent status switching, facilitating timely handling by management personnel and ensuring parking lot order. The overall design forms a complete closed loop from data acquisition and processing to status determination and anomaly warning, effectively improving parking lot management efficiency and intelligence.

[0025] In one embodiment, raw geomagnetic data and raw visual data of the target parking space in the parking lot are acquired and preprocessed separately, including geomagnetic data preprocessing: Obtain multi-dimensional raw data collected by the geomagnetic sensor of the target parking space:

[0026] in, This indicates the dimensions of geomagnetic data, including the x / y / z axis components of magnetic field strength. Indicates the data acquisition time step; The raw geomagnetic data is normalized to eliminate dimensional differences. The normalization expression is as follows:

[0027] in, Indicates the first 3D geomagnetic data in the first The original values ​​at each time step.

[0028] Indicates the first The mean of the geomagnetic data.

[0029] Indicates the first The standard deviation of geomagnetic data This is the normalized geomagnetic data.

[0030] This design preprocesses the raw geomagnetic data, first acquiring multi-dimensional raw data covering the components of magnetic field strength along each axis and time step information. Then, through normalization, the mean and standard deviation are used to eliminate dimensional differences. Geomagnetic data of different dimensions have different units and ranges, and direct use will affect the analysis results. Normalization makes the data within a similar range, making subsequent fusion and analysis more scientific and accurate. It can avoid deviations caused by differences in data dimensions, improve data usability, lay the foundation for accurately judging the parking space status, and enhance the system's adaptability to changes in different geomagnetic environments.

[0031] In one embodiment, the preprocessing further includes visual data preprocessing: High-definition cameras deployed above the parking spaces capture raw visual images of the target parking spaces. The images are then converted to grayscale, denoised, and have regions of interest (ROI) extracted to obtain images of the parking space area. Extract target features from the parking space area image and calculate feature parameters, including: Target contour area The Canny edge detection algorithm is used to extract the target contour in the image, and the pixel counting method is used to calculate the total number of pixels in the contour-enclosed region. ,in Indicates the first within the outline A valid pixel is identified by a value of 1, and an invalid pixel is identified by a value of 0. This represents the total number of pixels within the outline. Gray mean Calculate the average grayscale value of all pixels within the ROI region using the following formula:

[0032] in , These represent the width and height of the ROI region, respectively. Representing coordinates The grayscale value of the pixel at that location; Standardize the feature parameters into visual feature vectors:

[0033] in , , The threshold for the maximum area of ​​the preset parking space.

[0034] This design involves preprocessing the original visual image. First, the parking space area image is obtained through grayscale conversion, noise reduction, and ROI extraction. Then, feature parameters such as target contour area and grayscale mean are extracted. Finally, the image is standardized into a visual feature vector. The original image contains a lot of irrelevant information. Preprocessing can remove interference and highlight key features. The extracted feature parameters can accurately reflect the parking space status. Standardization makes different image features comparable, which helps to improve the quality of visual data, enhances the system's ability to recognize visual changes in parking spaces, and provides a reliable basis for subsequent fusion with geomagnetic data.

[0035] In one embodiment, a Kalman filter data fusion algorithm is used to fuse preprocessed geomagnetic data and visual data to obtain fused feature values ​​of parking space status, including: Construct the state equation and observation equation for the Kalman filter: Equations of state:

[0036] in, This represents the state vector at time t. This represents the parking space occupancy status value. The rate of change of state, Here is the state transition matrix. The sampling time interval, To control the input matrix, As a system control variable, set it to 0. For process noise, obey , The process noise covariance matrix; Observation equation:

[0037] in For the observation vector, Let be the mean of the geomagnetic data at time t. The weighted sum of the visual feature vectors at time t , These are the weighting coefficients. For the observation matrix, To observe noise, obey , To observe the noise covariance matrix; Kalman filter iterative calculation: Prediction step:

[0038] in The predicted state value at time t. The prediction error covariance matrix; Update steps: Calculate Kalman gain

[0039]

[0040] Status Update:

[0041] Error covariance update:

[0042] Extracting the updated state vector As a feature value for parking space status fusion.

[0043] This design utilizes the Kalman filter algorithm to fuse preprocessed geomagnetic and visual data, constructing state and observation equations to describe the relationship between system state and observation. Through iterative calculation, the state is predicted and updated, and the values ​​in the updated state vector are extracted as fused feature values. Single sensor data has limitations, but Kalman filtering can fuse the advantages of multi-source data, overcome noise interference, and through equation construction and iterative calculation, can more accurately estimate the parking space status, improve the reliability of fused feature values, and enable the system to comprehensively utilize geomagnetic and visual information to more accurately reflect the actual status of parking spaces, thereby improving monitoring accuracy.

[0044] In one embodiment, determining the occupancy status of a target parking space based on a comparison of the parking space status fusion feature value and a preset determination threshold includes: Preset judgment thresholds are calculated based on historical monitoring data. :

[0045] in This represents the mean of the fused feature values ​​when the parking space is vacant. The standard deviation of the fused feature values ​​in the idle state; Set the judgment rules: when At that time, the parking space is determined to be occupied; when At that time, the parking space is determined to be vacant; Abnormal condition determination: When continuous Each time step The parking space status conflicts with historical statistical patterns, such as continuous occupancy during idle periods and frequent switching of occupancy periods; Or the fusion feature value exceeds the normal range hour, The minimum effective threshold, The maximum effective threshold is used to determine an abnormal state.

[0046] This design determines parking space occupancy status based on fused feature values ​​and preset thresholds. First, the preset threshold is calculated, then the judgment rules are set, and rules for judging abnormal conditions are also set. The preset threshold is calculated based on historical data and can reflect the normal range of parking space status. The judgment rules are simple and clear, which can quickly and accurately determine the parking space occupancy status. The judgment of abnormal conditions can promptly detect abnormal parking space statuses, such as illegal occupation or frequent status switching, which helps managers to handle abnormalities in a timely manner, ensure parking lot order, and improve the system's practicality and intelligent management level.

[0047] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0048] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A parking lot occupancy monitoring and anomaly early warning system integrating geomagnetism and vision, characterized in that, The system includes: Data preprocessing unit, data fusion unit, status determination unit, and anomaly early warning unit; The data preprocessing unit is used to acquire and preprocess the geomagnetic raw data and visual raw data of the target parking space in the parking lot, respectively. The data fusion unit is used to perform fusion calculations on the preprocessed geomagnetic data and visual data using a Kalman filter data fusion algorithm to obtain fusion feature values ​​of parking space status. The status determination unit is used to determine the occupancy status of the target parking space based on the comparison between the parking space status fusion feature value and the preset determination threshold. The abnormal warning unit is used to trigger and output an abnormal warning signal when the parking space occupancy status meets the abnormal conditions.

2. The parking lot occupancy monitoring and anomaly early warning system integrating geomagnetism and vision as described in claim 1, characterized in that, The raw geomagnetic data is multi-dimensional data collected by geomagnetic sensors, including at least: The magnetic field strength has x-axis, y-axis, and z-axis components, with a data dimension of K and a data acquisition time step of T. The raw geomagnetic data is stored in the form of a K×T dimension matrix.

3. The parking lot occupancy monitoring and anomaly early warning system integrating geomagnetism and vision as described in claim 1, characterized in that, The geomagnetic data preprocessing includes: Normalization is performed to eliminate the dimensional differences between geomagnetic data of different dimensions, resulting in standardized geomagnetic data. The normalization process is calculated based on the mean and standard deviation of geomagnetic data in each dimension.

4. The parking space occupancy monitoring and anomaly early warning system based on geomagnetic and visual fusion as described in claim 1, characterized in that: The raw visual data is the target parking space image obtained by the image acquisition device deployed above the parking space; Visual data preprocessing includes image grayscale conversion, denoising, region of interest extraction, target feature extraction, and feature parameter standardization to obtain visual feature vectors.

5. The parking lot occupancy monitoring and anomaly early warning system based on geomagnetic and visual fusion as described in claim 4, characterized in that: The target features include at least the target contour area and the average gray level; The area of ​​the target contour is calculated by pixel counting after the contour is extracted by an edge detection algorithm. The mean gray level is the average gray level of all pixels within the region of interest.

6. The parking lot occupancy monitoring and anomaly early warning system integrating geomagnetism and vision according to claim 1, characterized in that, The Kalman filter data fusion algorithm includes: Construct the state equation and the observation equation; Data fusion is achieved through iterative calculations in the prediction and update steps; The state equation includes the parking space occupancy state value and the rate of change of the state; The observation vector of the observation equation is a weighted sum of the mean of geomagnetic data and the visual feature vector.

7. The parking lot occupancy monitoring and anomaly early warning system based on geomagnetic and visual fusion as described in claim 6, characterized in that: The prediction step calculates the predicted state value and the prediction error covariance matrix at time t; The update step involves calculating the Kalman gain, state update, and error covariance update. The parking space occupancy status value is extracted from the updated state vector and used as the parking space state fusion feature value.

8. The parking space occupancy monitoring and anomaly early warning system based on geomagnetic and visual fusion as described in claim 1, characterized in that: The preset judgment threshold is calculated based on the mean and standard deviation of the fused feature values ​​in the historical monitoring data of parking spaces in the vacancy state. By comparing the fused feature values ​​with the preset judgment threshold, the parking space is determined to be occupied or vacant.

9. The parking lot occupancy monitoring and anomaly early warning system integrating geomagnetism and vision according to claim 1, characterized in that, The abnormal conditions include: a first scenario and a second scenario: The first scenario is that within a continuously preset number of time steps, the parking space status conflicts with historical statistical patterns; The second scenario is when the fusion feature value of the parking space status exceeds the preset normal threshold range, and the preset number of time steps is not less than 5.

10. The parking lot occupancy monitoring and anomaly early warning system based on geomagnetic and visual fusion as described in claim 1, characterized in that, In the standardization process of the visual feature vector, the target contour area is normalized by comparing it with the contour area threshold of the maximum target that the preset parking space can accommodate, and the grayscale mean is normalized by comparing it with the maximum grayscale value.

Images