A sky-ground integrated cultural relic safety monitoring method and system

Abstract

The application provides a sky-ground integrated cultural relic safety monitoring method and system, relates to the technical field of cultural relic safety monitoring, and establishes a ground surface reference parameter by acquiring radar, unmanned aerial vehicle image data and ground vibration data of an acquisition device in a region of a target site at different time periods, combining geological exploration and meteorological data; compares the radar image to identify ground surface change information, generates a ground surface difference map marked with deviation values of each geographical position according to the deviation degree of the ground surface change information from the reference parameter; for key positions in the map that meet a preset deviation value, extracts corresponding unmanned aerial vehicle image slices, obtains visual feature vectors through multi-scale texture decomposition and color space conversion, combines ground vibration data to perform spatiotemporal correlation analysis and calculate a risk index, and determines that there is a violation activity to generate an alarm, so that accurate monitoring and rapid alarm of a violation activity of a cultural relic site can be realized.

Description

Technical Field

[0001] This application relates to the field of cultural relic safety monitoring technology, and in particular to a sky-ground integrated method and system for cultural relic safety monitoring. Background Technology

[0002] Cultural relics are irreplaceable and precious cultural heritage, and their safety and protection are of paramount importance. Monitoring the safety of cultural relics is a core means of preventing damage risks and achieving preventative protection, providing data support for cultural relic protection decisions. With the upgrading of cultural relic protection needs, integrated air-ground monitoring, due to its advantages of wide coverage and high monitoring efficiency, is gradually becoming the mainstream direction in the industry, with broad application prospects.

[0003] A new integrated air-ground method for monitoring cultural relic safety has been developed. This method acquires image data from different periods using satellite remote sensing and drone aerial photography, combines this data with environmental parameters collected by ground sensors, and identifies surface changes through simple image comparison. An alarm is automatically issued when the changes exceed a preset threshold. This method has already been implemented in the monitoring of some large-scale cultural relic sites, effectively compensating for the insufficient coverage of traditional manual inspections.

[0004] However, this existing method only focuses on comparing surface differences in images, failing to fully explore the inherent correlations between multi-source data. It lacks sufficient sensitivity in identifying subtle surface changes and struggles to accurately distinguish between changes caused by natural environmental evolution and those resulting from illegal activities. This leads to numerous false alarms and missed alarms, making it impossible to reliably determine illegal activities. Therefore, existing technologies suffer from insufficient accuracy in monitoring illegal activities related to cultural relics. Summary of the Invention

[0005] The purpose of this application is to provide an integrated air-ground method and system for monitoring the safety of cultural relics, in order to solve the problem of insufficient accuracy in monitoring illegal activities related to cultural relics in existing technologies.

[0006] To address the aforementioned technical problems, firstly, this application provides an integrated air-ground method for monitoring the safety of cultural relics, comprising:

[0007] The system acquires radar imagery data and UAV imagery data of the target site area at different time periods, as well as ground vibration data collected by acquisition devices deployed within the target site area.

[0008] Based on the geological survey data and meteorological data of the target site area, establish the surface benchmark parameters of the target site area;

[0009] By comparing radar image data from different time periods, surface change information of the target site area can be identified;

[0010] Based on the degree of deviation between the surface change information and the surface reference parameters, a surface difference map of the target site area is generated, which is used to represent the deviation value of each geographical location in the target site area.

[0011] For key geographical locations that meet the preset deviation value in the surface difference map, image slices corresponding to the key geographical locations are extracted from the UAV image data, and multi-scale texture decomposition and color space conversion are performed on the image slices to obtain the visual feature vectors of the image slices.

[0012] By combining the ground vibration data, a spatiotemporal correlation analysis is performed on the key geographical location and the visual feature vector to calculate the risk index of the key geographical location;

[0013] Based on the risk index, if it is determined that there are illegal activities within the target site area, an alarm event will be generated.

[0014] Optionally, establishing the surface reference parameters of the target site area based on the geological survey data and meteorological data of the target site area includes:

[0015] The target archaeological site area is divided into multiple geographical locations;

[0016] Based on the geological survey data, the geographical locations with the same geological attributes are classified into the corresponding geological zones;

[0017] Historical meteorological data of the target site area were obtained, and the correlation between the historical meteorological data and the surface manifestations of various geological zones in historical image data was analyzed in order to construct a quantitative model reflecting the impact of meteorological changes on different geological zones.

[0018] Based on real-time meteorological data at the time of image data acquisition for comparison, the corresponding benchmark value is calculated for each geological zone through the quantification model, and the benchmark value is assigned to all geographical locations within the geological zone to form a surface benchmark parameter covering each geographical location.

[0019] Optionally, the step of combining the ground vibration data to perform spatiotemporal correlation analysis on the key geographical location and the visual feature vector to calculate the risk index of the key geographical location includes:

[0020] The occurrence time and spatial source of the vibration event are determined based on the ground vibration data;

[0021] Using the occurrence time and spatial source of the vibration event as spatiotemporal constraints, a correlation analysis is performed on the key geographical locations in the surface difference map and the image slices corresponding to the visual feature vectors to calculate the risk index of the key geographical locations.

[0022] Optionally, the visual feature vector includes soil roughness, shadow sharpness, and deposit regularity;

[0023] Perform multi-scale texture decomposition and color space transformation on the image slices to obtain the visual feature vectors of the image slices, including:

[0024] The pixel color representation of the image slice is converted into a representation in which the luminance component and the color component are separated;

[0025] The transformed image slices are decomposed into multi-level layers to separate the basic information layer representing the overall contour and multiple detail information layers representing different levels of refinement.

[0026] Based on the multiple detailed information layers, soil roughness is calculated by analyzing the distribution intensity of high-frequency information;

[0027] Based on the brightness components, the shadow sharpness is calculated by quantifying the rate of brightness change at the edges of the identified low-brightness areas.

[0028] The regularity of the accumulation is calculated by analyzing the matching degree between the object contours identified in the image slices and the standard geometric shape, as well as the tortuosity of the contours themselves.

[0029] The calculation results of soil roughness, shadow sharpness, and deposit regularity are combined to form a visual feature vector.

[0030] Optionally, the visual feature vector includes soil roughness, shadow sharpness, and deposit regularity;

[0031] Using the occurrence time and spatial source point of the vibration event as spatiotemporal constraints, a correlation analysis is performed on key geographical locations in the surface difference map and the image slices corresponding to the visual feature vectors to calculate the risk index of the key geographical locations, including:

[0032] Based on the occurrence time and spatial source point of the vibration event, determine the time window and spatial radius;

[0033] Within the stated time window and spatial radius, key geographical locations are selected.

[0034] A base risk score is determined based on the deviation value of each of the key geographical locations;

[0035] The basic risk score is adjusted based on the visual feature vector associated with the key geographical location to obtain an adjusted risk score. The adjustment process includes increasing the basic risk score if the soil roughness, shadow sharpness, and deposit regularity meet preset conditions.

[0036] The risk index is calculated based on the adjusted risk score and the vibration event level extracted from ground motion data.

[0037] Optionally, the comparison of radar image data within different time periods to identify surface change information in the target site area includes:

[0038] Differential interferometry is used to process radar image data from different time periods to generate information on surface changes.

[0039] Optionally, it also includes:

[0040] The key geographical location is marked on the digital map interface corresponding to the target site area, and the radar image data, the UAV image data and the risk index used for the determination are displayed in conjunction.

[0041] Optionally, it also includes:

[0042] An online process for handling violations is initiated, which includes the steps of preliminary identification of activities at the key geographical locations, field evidence collection, legality determination, and field verification.

[0043] Update the processing status of the violation based on the progress of the online processing procedure.

[0044] Optionally, it also includes:

[0045] Based on the key geographical locations and risk indices, field verification tasks are generated and distributed to the mobile terminals of field personnel.

[0046] Receive and record the verification results uploaded by the mobile terminal, which include on-site text, images, or video evidence;

[0047] Update the status of the online processing procedure based on the on-site verification results.

[0048] Secondly, this application provides an integrated air-ground cultural relic security monitoring system, including:

[0049] The acquisition module is used to acquire radar image data, UAV image data and ground vibration data of the target site area at different time periods.

[0050] A module is established to establish surface reference parameters for the target site area based on geological survey data and meteorological data of the target site area.

[0051] The identification module is used to compare radar image data within different time periods to identify surface change information in the target site area;

[0052] The first generation module is used to generate a surface difference map of the target site area based on the degree of deviation between the surface change information and the surface reference parameters. The surface difference map is used to represent the deviation values ​​of each geographical location in the target site area.

[0053] The extraction module is used to extract image slices corresponding to key geographical locations that meet preset deviation values ​​in the surface difference map from the UAV image data, and to perform multi-scale texture decomposition and color space conversion on the image slices to obtain the visual feature vectors of the image slices.

[0054] The analysis module is used to combine the ground vibration data to perform spatiotemporal correlation analysis on the key geographical location and the visual feature vector, so as to calculate the risk index of the key geographical location.

[0055] The second generation module is used to generate an alarm event if it is determined that there is illegal activity within the target site area based on the risk index.

[0056] This application provides an integrated air-ground cultural relic safety monitoring method that provides comprehensive data support for monitoring by acquiring multi-source data; establishes surface benchmark parameters to clarify the reference for the normal state of the surface; accurately identifies surface changes by comparing radar images; generates surface difference maps to intuitively present the degree of deviation of each geographical location; extracts visual feature vectors from image slices to capture key location details; combines ground vibration data with spatiotemporal correlation analysis to accurately calculate the risk level; and determines violations and generates alarms to achieve timely risk warnings.

[0057] Furthermore, the target site area is divided into multiple geographical locations. Based on geological survey data, geographical locations with similar geological attributes are categorized into corresponding geological zones. Historical meteorological data for the region is then acquired and analyzed to correlate with the surface manifestations of each geological zone in historical imagery. A quantitative model of the impact of meteorological changes on different geological zones is constructed. Finally, based on real-time meteorological data at the time of image acquisition, benchmark values ​​are calculated for each geological zone using the quantitative model and assigned to all geographical locations within the area, forming surface benchmark parameters covering each geographical location. This ensures that the surface benchmark parameters accurately reflect the actual geological and meteorological conditions of different regions, providing comprehensive coverage and high accuracy. This provides a reliable and realistic reference for accurately assessing the deviation between surface changes and benchmark parameters. Attached Figure Description

[0058] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 A schematic flowchart illustrating an integrated air-ground cultural relic security monitoring method provided in this application embodiment;

[0060] Figure 2 A flowchart illustrating another integrated air-ground cultural relic security monitoring method provided in this application embodiment;

[0061] Figure 3 A flowchart illustrating another integrated air-ground cultural relic security monitoring method provided in this application embodiment;

[0062] Figure 4 This is a schematic diagram of a sky-ground integrated cultural relic safety monitoring system provided in an embodiment of this application. Detailed Implementation

[0063] Existing integrated air-ground cultural relic security monitoring methods only perform surface comparisons of image data from different periods, failing to fully integrate the intrinsic relationships between various data sources such as radar images, ground vibrations, geology, and meteorology. This makes it difficult for the method to accurately distinguish between natural environmental evolution and surface changes caused by human violations. Furthermore, it lacks sufficient sensitivity to identify subtle surface alterations, ultimately leading to numerous false alarms and missed alarms during monitoring. Consequently, it cannot reliably determine violations, severely impacting the timeliness and effectiveness of cultural relic protection.

[0064] To address the aforementioned issues, this invention proposes an integrated air-ground method for monitoring cultural relic safety. The core of this method lies in comprehensively integrating multi-source data and conducting in-depth correlation analysis. First, it collects radar, UAV imagery, and ground seismic data from different times, combining this with geological and meteorological information to establish a surface benchmark that accurately reflects the site's condition. Then, it identifies surface changes through image comparison, focusing on key change locations to extract detailed features. Finally, it calculates a risk index based on spatiotemporal correlation using seismic data, generating an alarm upon determining a violation. This approach overcomes the limitations of existing technologies that rely solely on surface image comparison. By collaboratively verifying multiple data sources, it accurately distinguishes change types and captures subtle anomalies, fundamentally solving the problems of insufficient accuracy and numerous false alarms and missed alarms in existing technologies. This significantly improves the reliability and timeliness of monitoring violations related to cultural relics, providing stronger technical support for cultural relic protection.

[0065] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0066] The core data for each dimension of sky, airspace, and ground in this application are specifically composed as follows:

[0067] Tianweidu: The core data is radar imagery data;

[0068] Aerial dimension: The core data is drone imagery data;

[0069] Geographic dimension: Core data include ground motion data, geological survey data, and meteorological data.

[0070] The methods for collecting data from each dimension and their core uses will be explained in further detail in subsequent steps and implementation details.

[0071] The core of this application is to provide an integrated air-ground method for monitoring the safety of cultural relics, and a flowchart of one specific implementation method is shown below. Figure 1 As shown, the method includes:

[0072] S101. Acquire radar image data, UAV image data, and ground vibration data collected by acquisition devices deployed in the target site area at different time periods.

[0073] Among them, radar image data refers to high-altitude image data obtained through satellite remote sensing technology, which can comprehensively reflect the overall surface conditions of the site area and capture large-scale macro changes, such as newly added illegal buildings and overall terrain disturbances.

[0074] Among them, drone image data refers to high-definition image data taken by drones when flying at low altitudes. It is good at focusing on details of local areas, such as subtle changes in the ground surface or small accumulations in a certain area.

[0075] Among them, the data acquisition device is a monitoring device specifically deployed in the target site area, including vibration sensors, ground sound detectors, accelerometers, etc., used to capture vibration signals generated by ground activities such as construction, blasting, and mechanical operations in real time; ground vibration data is ground vibration information recorded by the data acquisition device, which can indirectly reflect whether there are human activities around the site that may affect the safety of cultural relics.

[0076] In one specific implementation, radar image data at different times is acquired by periodic satellite photography, covering the entire target site area. The photography cycle can be quarterly or semi-annually, without specific limitations. Drones are used to take aerial photos at low altitudes along preset routes to acquire drone image data within the corresponding time period, focusing on areas where anomalies may exist. Ground vibration data is continuously collected and recorded by data acquisition devices pre-deployed at key locations of the site, such as the boundaries of the protection area and around core cultural relics.

[0077] S102. Based on the geological survey data and meteorological data of the target site area, establish the surface reference parameters of the target site area.

[0078] Among them, geological survey data includes information on the stratigraphic structure, soil type, and foundation stability of the archaeological site area, obtained through geological exploration equipment. Meteorological data includes historical and real-time meteorological data, such as records of rainfall, sandstorms, and temperatures over the years, as well as weather conditions at the time of image acquisition.

[0079] S102 specifically includes:

[0080] S1021. Divide the target site area into multiple geographical locations.

[0081] S1022. Based on the geological survey data, the geographical locations with the same geological attributes are classified into the corresponding geological zones.

[0082] S1023. Obtain historical meteorological data of the target site area and analyze the correlation between the historical meteorological data and the surface manifestations of various geological zones in the historical image data, so as to construct a quantitative model reflecting the impact of meteorological changes on different geological zones.

[0083] Among them, the quantitative model is a mathematical model that quantifies the degree of impact of meteorological changes on the surface of different geological zones, and is used to calculate a benchmark value that fits reality.

[0084] S1024. Based on the real-time meteorological data at the time of image data acquisition for comparison, the corresponding benchmark value is calculated for each geological zone through the quantification model, and the benchmark value is assigned to all geographical locations within the geological zone to form a surface benchmark parameter covering each geographical location.

[0085] Among them, the surface benchmark parameters are reference standards that reflect the normal surface state of various geographical locations in the target site area, and are used to subsequently determine whether abnormal changes have occurred on the surface.

[0086] In one specific implementation, such as Figure 2As shown, the target site area is first divided into multiple geographical locations. Then, based on geological survey data, geographical locations with the same geological attributes are classified into corresponding geological zones. Next, historical meteorological data of the target site area is obtained and its correlation with the surface manifestations of various geological zones in historical image data is analyzed to construct a quantitative model. Finally, based on real-time meteorological data at the time of image data acquisition, the quantitative model is used to calculate the benchmark value for each geological zone and assign it to all geographical locations within the zone. Combining geological and meteorological characteristics, the benchmark parameters for the entire surface area are established.

[0087] As an example, consider an ancient city site that includes both earthen and stone inscription areas:

[0088] First, in step S1021, the ancient city site is divided into multiple independent geographical locations with an area of ​​10 meters × 10 meters. Each location is assigned a unique identifier, such as A1, A2…An, to facilitate precise management later. The above example is only one example of this application. In practical applications, the area division can be adjusted according to the size of the site and the complexity of the terrain. This application does not limit this.

[0089] Secondly, in step S1022, geological survey data, such as stratigraphic structure data obtained from geological radar and soil type data obtained from soil sample analysis, are used to classify geographical locations with the same geological attributes into different geological zones. For example, A1-A100, with loess soil and similar foundation stability, are classified as earthen site zones, while A101-A200, with rock texture and hard structure, are classified as stone cultural relics zones.

[0090] Next, in step S1023, historical meteorological data of the ancient city site over the past 30 years were obtained, such as annual total rainfall, number of sandstorm days, number of extreme temperature events, and historical remote sensing image data for the corresponding years. The correlation between the two types of data was analyzed: the degree of surface weathering in the earthen site area when the annual rainfall was 500 mm was statistically analyzed. In this example, the degree of surface weathering was reflected by the number of surface cracks and the area of ​​peeling in the historical images. The surface erosion was also statistically analyzed when there were more than 30 sandstorm days. The changes in surface weathering traces in the stone cultural relic area were statistically analyzed when there were more than 20 days of extreme high temperature (≥35℃).

[0091] A quantitative model was then constructed using a multiple linear regression algorithm. The dependent variable was an index of normal land surface condition, and the independent variables were core meteorological factors. The dependent variable, denoted as S, represents the surface smoothness by the pixel grayscale variance at that location in historical images, with a value ranging from 0 to 100. Higher values ​​indicate smoother and more normal land surface conditions. The independent variables were annual rainfall R, number of windy and sandy days F, and number of extreme temperature days T. The coefficients of the independent variables varied for different geological zones. The model formula is as follows:

[0092] Earthen site zoning: (1);

[0093] Stone artifact zoning: (2);

[0094] In the formula, , , , The regression coefficients for the earthen site zoning are given. , , , Here, R represents the annual rainfall (mm), F represents the number of days with sandstorms (days), and T represents the number of days with extreme temperatures (days). , This is the benchmark value for the surface flatness of the corresponding zone.

[0095] During model training, 20 years of historical data from the past 30 years are selected as the training set. The R, F, and T values ​​for each year are substituted into the formula, and the least squares method is used to minimize the error between the predicted value and the actual S value extracted from historical imagery, thus solving for the coefficients. For example, after training on earthen site areas, the following results are obtained: , , , Stone cultural relics were zoned. , , , The remaining 10 years of data will be used as a test set to verify that the model's prediction error is within 5%, ensuring the model's effectiveness.

[0096] In another specific implementation, a random forest algorithm can be used to construct a quantitative model. The prediction accuracy can be improved by ensemble learning of multiple decision trees. The independent variables can include meteorological factors such as humidity and the number of lightning strikes. This application does not limit the model algorithm and the selection of independent variables.

[0097] Finally, real-time meteorological data at the time of radar image acquisition is obtained through step S1024. At the time of acquisition, the real-time rainfall in the area is R=50mm, the number of windy and sandy days is F=5 days, and the number of extreme temperature days is T=3 days. These data are then substituted into the quantitative model of the corresponding geological zone to calculate the baseline values, i.e., substituted into equation (1) or equation (2):

[0098] Earthen site zoning benchmark values: ;

[0099] Standard values ​​for zoning stone cultural relics: ;

[0100] The calculated benchmark values ​​are assigned to all geographical locations within the corresponding geological zones. For example, the benchmark values ​​for A1-A100 are all 86.28, and the benchmark values ​​for A101-A200 are all 90.71, thus forming surface benchmark parameters covering each geographical location of the ancient city site.

[0101] This application constructs a quantitative model by combining geological zoning and meteorological influences, so that the surface benchmark parameters are consistent with the actual conditions of different regions, accurately reflecting the normal surface state of each geographical location, and providing a reliable and personalized reference for subsequent accurate judgment of whether surface changes are abnormal.

[0102] S103. Compare radar image data within different time periods to identify surface change information in the target site area.

[0103] S103 specifically includes:

[0104] Differential interferometry is used to process radar image data from different time periods to generate information on surface changes.

[0105] Differential interferometry (DI) is a technique that processes radar images of the same area acquired at different times to calculate minute surface deformations and positional changes, enabling precise capture of subtle surface alterations. Surface change information specifically refers to changes in the surface condition of archaeological sites caused by human activities, including changes in surface morphology resulting from illegal construction, the accumulation of construction materials, and excavation.

[0106] In one specific implementation, this step first selects radar image data of the target site area at different time periods, processes the images using differential interferometry, obtains surface change data by calculating the phase difference between the images, and then identifies surface change information caused by human activities.

[0107] As an example, taking the previously mentioned ancient city ruins as an example, we first acquired two sets of radar image data for the first and third quarters of a certain year. Both sets of data completely covered the entire ancient city ruins area, ensuring that no monitoring range was missed. Next, we preprocessed the two sets of images, including image registration and noise removal, so that the same geographical location corresponds accurately in the two sets of images and reduces irrelevant interference factors.

[0108] Subsequently, differential interferometry is used to calculate surface changes. The core principle is to infer surface changes by using phase differences. The phase difference calculation formula is as follows:

[0109] (3)

[0110] In the formula, The effective phase difference corresponding to surface changes, with a value range of 0-2π; φ1 represents the pixel phase value of the image in the next time period (third quarter), while φ2 represents the pixel phase value of the image in the previous time period (first quarter). Both are the original phase information inherent in the radar images.

[0111] Taking location A50 of the ancient city ruins, an open space within the earthen ruins section, as an example, its corresponding pixel... , Substitute into equation (3) to calculate: .

[0112] Then, based on the conversion relationship between phase difference and surface deformation, the actual surface deformation is further calculated. The conversion formula is as follows:

[0113] (4)

[0114] In the formula, The value represents the surface deformation, in meters (m). The wavelength of the radar image is used in this example. π is the mathematical constant of a circle, with a value of 3.14.

[0115] Will Substitution formula (4): .

[0116] By performing the above calculations on each pixel in the two sets of images, the surface deformation of each geographical location of the entire ancient city site is obtained, which is used as the surface change information output in this step.

[0117] The above example is only one example of this application and is applicable to site monitoring scenarios with flat terrain and stable atmospheric conditions. In practical applications, if the site terrain is complex or atmospheric interference is obvious, terrain and atmospheric correction steps can be added as needed. The time period and wavelength selection of radar images can also be flexibly adjusted. This application does not limit this.

[0118] In another specific implementation, long-term remote sensing image comparison analysis can be used to identify surface change information. This method constructs a complete time series by selecting radar images from multiple consecutive time periods. Relying on pixel-level feature comparison and trend analysis, it accurately calculates the surface deformation of each geographical location, effectively eliminating temporary surface fluctuations caused by short-term natural factors. It is especially suitable for monitoring long-term covert illegal activities. This application does not limit the specific type of interferometry technology used.

[0119] This application uses differential interferometry to process radar images from different time periods, which can accurately calculate the surface deformation of each geographical location, providing accurate basic data on surface changes for subsequent comparison with baseline parameters and generation of difference maps.

[0120] S104. Based on the degree of deviation between the surface change information and the surface reference parameters, generate a surface difference map of the target site area. The surface difference map is used to represent the deviation values ​​of each geographical location in the target site area.

[0121] The deviation value is the difference between the actual surface change at each geographical location within the target site area and the corresponding surface baseline parameter, used to quantify the degree to which the surface deviates from the normal state. The surface difference map is a visual chart integrating all geographical deviation values, which can intuitively present the deviation at each geographical location and facilitate the quick identification of abnormal areas with significant deviations.

[0122] In one specific implementation, the surface deformation of each geographical location obtained in S103 is first matched one by one with the corresponding surface reference parameters established in S102, the deviation value of each geographical location is calculated, and then the deviation value is visualized and marked using a digital map as the base map to generate a surface difference map. At the same time, key geographical locations that meet the preset deviation value are selected.

[0123] As an example, taking the ancient city ruins as an example, first retrieve the surface deformation variables of each geographical location calculated in step S103, such as Δh=0.01159m at location A50, and the corresponding surface reference parameters established in step S102. Location A50 belongs to the earthen ruins zone, with a reference value S1=86.28 and a corresponding surface deformation reference threshold of 0.003m, or 0.3cm. This threshold is set by the reference parameters in combination with historical natural evolution data.

[0124] Next, the deviation value for each geographical location is calculated. The formula for calculating the deviation value is as follows:

[0125] (5)

[0126] In the formula, This is the deviation value, in meters (m). The actual surface deformation obtained from S103; The reference threshold for surface deformation at the corresponding location is derived from the surface reference parameters and reflects the reasonable range of deformation due to natural evolution.

[0127] Taking position A50 as an example, =0.003m, substituting into equation (5) yields... By calculating the deviation value of each geographical location of the ancient city ruins, the deviation value of each geographical location was obtained.

[0128] A surface difference map was then generated, using the digital map of the ancient city ruins as the base map. The deviation value of each geographical location was marked in two ways: first, the specific deviation value was marked at the corresponding location; second, a color gradient was used to distinguish the degree of deviation, with the larger the deviation, the more conspicuous the corresponding color. The deviation value of location A50 is 0.00859m, corresponding to red. Multiple consecutive pixels around it with deviation values ​​greater than 0.008m are also marked in red.

[0129] Finally, a preset deviation value is set, which is 0.003m in this example. Geographical locations with deviation values ​​exceeding this preset value, such as A50, A49, A51, and A52, are selected as key geographical locations to provide target areas for subsequent extraction of image slices.

[0130] The above example is only one example of this application. In practical applications, the method of calculating deviation value, the form of surface difference chart annotation, and the preset deviation value size can all be adjusted according to the type of site and the monitoring accuracy requirements. This application does not limit these aspects.

[0131] This application accurately quantifies the degree of deviation of each geographical location by comparing surface change information with benchmark parameters and generates a visual difference map, which can quickly identify key geographical locations, focus targets for subsequent feature extraction and risk analysis, and improve monitoring efficiency and targeting.

[0132] S105. For key geographical locations that meet the preset deviation value in the surface difference map, extract image slices corresponding to the key geographical locations from the UAV image data, and perform multi-scale texture decomposition and color space conversion on the image slices to obtain the visual feature vectors of the image slices.

[0133] Multi-scale texture decomposition is a processing technique that breaks down an image into parts of varying fineness to separate the overall contour from subtle texture information. Color space conversion is a method that separates the color representation of an image pixel into brightness and chroma to prevent them from interfering with each other and affecting feature extraction. Visual feature vectors are comprehensive data that integrates soil roughness, shadow sharpness, and accumulation regularity to quantify and describe the detailed features of image slices.

[0134] In step S105, multi-scale texture decomposition and color space conversion are performed on the image slices to obtain the visual feature vectors of the image slices, specifically including:

[0135] S1051. Convert the pixel color representation of the image slice into a representation in which the luminance component and color component are separated.

[0136] S1052. Perform multi-level decomposition on the converted image slices to separate the basic information layer representing the overall outline and multiple detail information layers representing different levels of refinement.

[0137] S1053. Based on the multiple detailed information layers, the soil roughness is calculated by analyzing the distribution intensity of high-frequency information.

[0138] Among them, soil roughness is an indicator that reflects the roughness of the surface soil and can be associated with traces of human activities such as construction and excavation.

[0139] S1054. Based on the brightness component, the shadow sharpness is calculated by quantifying the rate of brightness change at the edges of the identified low-brightness areas.

[0140] Among them, shadow sharpness is an indicator that measures the clarity of shadow edges and can identify shadow features formed by newly added objects.

[0141] S1055. By analyzing the matching degree between the object contours identified in the image slices and the standard geometric shape, as well as the curvature of the contours themselves, the regularity of the accumulation is calculated.

[0142] Among them, the regularity of the deposit is an indicator for evaluating the regularity of the shape of the deposited object, and is used to distinguish between natural deposits and man-made deposits.

[0143] S1056. Combine the calculation results of soil roughness, shadow sharpness and deposit regularity to form a visual feature vector.

[0144] In one specific implementation, this step first filters key geographical locations and extracts corresponding UAV image slices, then splits the image information through color space conversion and multi-level decomposition, and then calculates three visual features respectively, finally combining them to form a visual feature vector.

[0145] As an example, taking the ancient city ruins as an example, the key geographical locations with deviation values ​​exceeding the preset value are first extracted from the surface difference map, namely A50, A49, A51, and A52. Focusing on the core area A50, image slices of this area are extracted from the drone imagery data of the third quarter of 2023, with a size of 500 pixels × 500 pixels, to ensure that the slices completely cover the key areas.

[0146] Next, color space conversion is performed in step S1051, using RGB to YUV conversion. The color representation of a pixel is decomposed into the luminance component Y and the color components U and V. The conversion formula is as follows:

[0147] In the formula, R, G, and B are the red, green, and blue channel values ​​of the original image pixels, respectively, with values ​​ranging from 0 to 255; Y is the luminance component, with a value ranging from 0 to 255; and U and V are the color components, with values ​​ranging from -128 to 127.

[0148] (6)

[0149] (7)

[0150] (8)

[0151] Taking a pixel in an image slice as an example, with R=180, G=170, and B=160, substituting into equations (6), (7), and (8) yields: , , .

[0152] By calculating each pixel in the slice, we obtain the Y component image containing only brightness information and the U and V component images containing only color information. Subsequent feature extraction focuses on the Y component image.

[0153] Next, multi-level decomposition is performed in step S1052. The Laplacian pyramid algorithm is used to decompose the Y component image into 4 levels: the 0th level is the basic information layer, which represents the overall outline, such as large open areas and obvious building outlines; the 1st to 3rd levels are the detail information layers, which represent three levels of fine texture, such as soil particles, minor depressions, and object edge textures. The decomposition process is achieved through Gaussian blur and image subtraction. The size of each layer image is consistent with the original Y component image.

[0154] Then, soil roughness is calculated in step S1053. Based on the detail information layers 1-3, high-frequency information of each layer is extracted (high-frequency information corresponds to areas in the image where gray values ​​change rapidly, i.e., rough texture areas). The distribution intensity is quantized using the variance of high-frequency information, as shown in the following formula:

[0155] (9)

[0156] In the formula, R is the soil roughness, ranging from 0 to 100, with higher values ​​indicating rougher soil; i is the detail information layer number (1-3); M and N are the width and height of the detail information layer, respectively, and in this example, M=N=500; The high-frequency information value of the (x, y) coordinates in the i-th detail information layer; is the average value of the high-frequency information of the i-th layer.

[0157] The first layer was calculated. =25, variance is 120, second layer =18, variance is 95, 3rd layer =12, variance is 70, substituting into equation (9) yields soil roughness R as 95.

[0158] Subsequently, shadow sharpness is calculated in step S1054. Low-brightness areas are identified based on the Y component image, and areas with Y < 80 are defined as shadow areas. The edge contours of the shadow areas are extracted, and the rate of brightness change at the edges is calculated using gradient magnitude quantization, as shown in the following formula:

[0159] (10)

[0160] In the formula, S represents the shadow sharpness, with a value ranging from 0 to 50. A higher value indicates a sharper shadow edge; K represents the total number of pixels at the shadow edge. , These are the brightness gradients of the k-th edge pixel in the x and y directions, respectively, calculated using the Sobel operator.

[0161] After identification, the total number of shadow edge pixels of the image slice at position A50 is K=230. After calculating the gradient magnitude of each edge pixel and summing them, the total gradient is 5750. Substituting into equation (10), the shadow sharpness S is 25.

[0162] Next, the regularity of the accumulation is calculated in step S1055, using a weighted average of the contour matching degree and the contour tortuosity, as shown in the following formula:

[0163] (11)

[0164] In the formula, D represents the regularity of the accumulation, ranging from 0 to 100, with higher values ​​indicating greater regularity; M represents the matching degree between the object's outline and the standard geometric shape, calculated using the shape moment algorithm, ranging from 0 to 100; C represents the total tortuous length of the outline; and L represents the straight-line distance of the outline. The value range is 1-∞, where 1 indicates that the contour is a straight line.

[0165] Image recognition revealed an accumulation at position A50. The matching degree between its outline and the rectangle was M=85, the total tortuous length of the outline was C=120 pixels, and the straight-line distance of the outline was L=100 pixels. Substituting these values ​​into equation (11), the regularity of the accumulation was calculated as follows: The result is 51, rounded to the nearest whole number.

[0166] Finally, by combining the calculation results of the three features in step S1056, the soil roughness R=95, shadow sharpness S=25, and accumulation regularity D=51 are arranged in order to form the visual feature vector [95, 25, 51] of the image slice.

[0167] The above example is only one example of this application. In practical applications, the image slice size, color space conversion method, number of decomposition levels, feature calculation weights, etc. can all be adjusted according to the type of site and the monitoring accuracy requirements. This application does not limit these aspects.

[0168] In another specific implementation, the wavelet transform algorithm can be used instead of the Laplacian pyramid algorithm for multi-scale texture decomposition. This algorithm can more accurately preserve texture details at different scales. Visual features can also be enhanced by adding indicators such as color uniformity and object density to further enrich the feature dimensions and improve the accuracy of subsequent risk assessment. This application does not limit the selection of specific algorithms and feature indicators.

[0169] This application uses multi-step image processing and feature extraction to accurately capture visual details of key geographical locations, providing concrete and quantifiable visual evidence for subsequent identification of illegal activities based on ground seismic data, thereby improving the targeting and reliability of feature recognition.

[0170] S106. Combining the ground vibration data, perform spatiotemporal correlation analysis on the key geographical location and the visual feature vector to calculate the risk index of the key geographical location.

[0171] Spatiotemporal correlation analysis is an analytical method that matches the spatial information of key geographical locations, the detailed feature information of visual feature vectors, and the temporal and spatial information of ground vibration data to determine whether the three originate from the same human activity.

[0172] S106 specifically includes:

[0173] S1061. Determine the occurrence time and spatial source of the vibration event based on the ground vibration data.

[0174] Among them, the time of occurrence of the vibration event is the specific moment when a significant vibration signal is captured in the ground motion data, and the spatial source point is the specific geographical location that caused the vibration.

[0175] S1062. Using the occurrence time and spatial source of the vibration event as spatiotemporal constraints, perform correlation analysis on the key geographical locations in the surface difference map and the image slices corresponding to the visual feature vectors to calculate the risk index of the key geographical locations.

[0176] S1062 specifically includes:

[0177] Based on the occurrence time and spatial source of the vibration event, a time window and spatial radius are determined; key geographical locations are selected within the time window and spatial radius; a basic risk score is determined based on the deviation value of each key geographical location; the basic risk score is adjusted based on the visual feature vector associated with the key geographical location to obtain an adjusted risk score, wherein the adjustment process includes: increasing the basic risk score if the soil roughness, shadow sharpness, and sediment regularity meet preset conditions; and calculating a risk index based on the adjusted risk score and the vibration event level extracted from the ground vibration data.

[0178] The time window is a specific duration range defined around the time of the vibration event, used to filter surface changes that may be related to the vibration. The spatial radius is a specific geographical range defined around the spatial source point of the vibration event, used to pinpoint key geographical locations within the vibration's impact area. The basic risk score is an initial risk score converted from the deviation value of key geographical locations, reflecting the degree of risk of the surface deviating from its normal state. The vibration event level is an intensity level classified according to the amplitude, frequency, and duration of the vibration, used to reflect the severity of the activity that caused the vibration.

[0179] In one specific implementation, this step first uses ground motion data to lock in the spatiotemporal information of vibration events, then uses this information as a constraint to filter relevant key geographical locations, calculates the basic risk score and the adjusted risk score in sequence, and finally combines the vibration event level to obtain the risk index, forming a complete risk quantification process.

[0180] As an example, let's take ancient city ruins as an example again:

[0181] First, the occurrence time and spatial source of the vibration event are determined through step S1061. A peak detection algorithm is used to analyze ground vibration data deployed at the ancient city ruins. A vibration amplitude threshold is set, such as 0.01 m / s². When a vibration amplitude exceeding the threshold is detected and the duration is ≥3 seconds, it is determined to be a valid vibration event, and its peak time is recorded as the occurrence time.

[0182] Simultaneously, the time difference of vibration arrival was collected using three vibration sensors distributed at different locations within the site. These sensors were designated S1, S2, and S3, and the spatial source point was calculated using a triangulation algorithm. For example, if S1 detected vibration at 14:30:20 on July 15th of a certain year, S2 at 14:30:22, and S3 at 14:30:21, and the coordinates of S1, S2, and S3 are known to be (100, 200), (150, 250), and (120, 230) respectively (in meters), and the vibration propagation speed is taken as 340 m / s (the speed of vibration propagation in the soil), the spatial source point coordinates were calculated using the triangulation formula to be (130, 220), corresponding to the vicinity of location A50 of the ancient city site. Location A50 of the ancient city site is a key geographical location.

[0183] Secondly, correlation analysis and risk index calculation are performed using 1062, which involves four steps:

[0184] The first step was to determine the time window and spatial radius. Based on experience in monitoring the site, the time window was set to one hour before and after the vibration occurred, and the spatial radius was set to 50 meters. Key geographical locations within the spatial radius were then selected from the surface difference map within this time window, namely location A50 and the surrounding four key geographical locations A49, A51, and A52.

[0185] The second step is to calculate the baseline risk score. Using a linear normalization method, the deviations from key geographical locations are converted into a baseline risk score ranging from 0 to 50. The calculation formula is as follows:

[0186] (12)

[0187] In the formula, Basic risk score (range 0-50 points); This represents the deviation from the key geographical location, in meters (m). This is the minimum of all selected key geographic location deviations; in this example, it is 0.005m. This represents the maximum value of all filtered key geographic location deviations; in this example, it is 0.00859m.

[0188] Deviation of A50 position =0.00859m, substituting into equation (12) yields...

[0189] The third step is to adjust the base risk score based on the visual feature vector. Preset conditions are established: soil roughness ≥ 80, shadow sharpness ≥ 20, and sediment regularity ≥ 50. Meeting any one of these conditions adds 5 points, with a maximum of 15 points. The visual feature vector at position A50 is [95, 25, 51]. All three conditions are met, therefore the adjusted risk score is [95, 25, 51]. =50+15=65 points.

[0190] The fourth step is to extract the vibration event level and calculate the risk index. Vibration events are classified into levels based on vibration amplitude: 0.01-0.02 m / s² is Level 1 (mild), 0.02-0.03 m / s² is Level 2 (moderate), and ≥0.03 m / s² is Level 3 (severe), with corresponding level coefficients K of 1.0, 1.2, and 1.5, respectively. In this example, the vibration event amplitude is 0.025 m / s², the level is 2, and the level coefficient K = 1.2. The adjusted risk score is then weighted according to the level coefficient to calculate the risk index, which is 78 points in this example.

[0191] The above example is only one example of this application. In practical applications, the duration of the time window, the size of the spatial radius, the deviation threshold, the preset conditions of visual features, the classification standards of vibration events, etc. can all be adjusted according to the type of site and the requirements of monitoring accuracy. This application does not limit these.

[0192] This application accurately reflects the degree of violation risk in key geographical locations through spatiotemporal correlation and quantitative calculation of multi-dimensional data, providing a reliable quantitative basis for subsequent violation judgment, reducing misjudgment and omission, and achieving deep integration between the data of the preceding steps and the risk calculation of this step.

[0193] S107. Based on the risk index, if it is determined that there are illegal activities within the target site area, an alarm event is generated.

[0194] The alert events are warning messages indicating suspected illegal activities within the archaeological site area. These messages include key geographical locations, risk indices, and relevant data sources, triggering subsequent handling procedures. The determination of illegal activities is based primarily on the risk index, using reasonable thresholds to differentiate between normal fluctuations and suspected violations, ensuring the accuracy of the alerts.

[0195] In one specific implementation, this step first sets a risk index threshold, compares the risk index of the key geographical location calculated in S106 with the threshold, determines whether there is any illegal activity, and if the threshold is met, generates an alarm event containing complete information and simultaneously triggers the subsequent handling process.

[0196] As an example, taking the ancient city ruins as an example, a risk index threshold of 60 points is first set. This threshold can be adjusted according to the site's protection level and historical violations. The calculated risk index for location A50 is 78 points, exceeding the threshold, indicating suspected illegal activities at and around location A50. An alarm event is then generated. The basic information of the event includes: administrative division, name of the cultural relic, protection level, and time-phase information (i.e., the time corresponding to the period of surface change information, the center point location, i.e., the key geographical location), the geographical location corresponding to A50, and the risk index. The generated alarm event is simultaneously pushed to the cultural relic safety monitoring platform, triggering subsequent map display and online processing procedures.

[0197] Following S107, it also includes:

[0198] The key geographical location is marked on the digital map interface corresponding to the target site area, and the radar image data, the UAV image data and the risk index used for the determination are displayed in conjunction.

[0199] The digital map interface is a visual interface within the cultural relic safety monitoring platform that integrates the site's geographical information and monitoring data. It supports geographic location marking and data association viewing, allowing managers to intuitively grasp the situation of areas with violations. The data association display binds various raw data used for alarm judgments with key geographic locations, ensuring that the basis for judgments is traceable and verifiable.

[0200] Initiate an online process for handling violations, which includes steps such as preliminary identification of activities at the key geographical locations, field evidence collection, legality determination, and field verification; update the processing status of the violations based on the progress of the online process.

[0201] Specifically, field evidence collection includes:

[0202] Based on the key geographical location and risk index, generate field verification tasks and distribute them to the mobile terminals of field personnel; receive and record verification results uploaded by the mobile terminals, including on-site text, images, or video evidence; and update the status of the online processing procedure based on the on-site verification results.

[0203] Field evidence collection is the core of the online processing procedure. It involves collecting evidence on-site to verify suspected violations and provide objective evidence for legality determination. Field verification tasks are standardized task sheets that include key geographical locations, risk indices, suspected violation types, and evidence requirements. Mobile terminals are smart devices carried by field personnel, such as smartphones and tablets, which have functions such as positioning, taking photos, recording videos, text input, and data uploading, realizing a closed loop of task reception and result feedback.

[0204] In one specific implementation, the illegal areas are first visually marked and linked to data. Then, the process is connected online to link the initial judgment, field evidence collection, legality determination and field verification stages, and the progress is tracked throughout.

[0205] As an example, following the previous alarm incident at the ancient city ruins, the key geographical locations are first marked on the digital map interface of the cultural relic safety monitoring platform: the core area of ​​location A50 is marked with a red highlight, and the surrounding related areas are marked with a colored box. Clicking on the marked point will pop up a details window, which displays three core data: radar image data, drone image data, and risk index.

[0206] Then an online processing procedure was initiated, such as Figure 3 As shown, the specific steps are as follows:

[0207] The first step was a preliminary assessment. Based on the data linked to the map interface, the management personnel initially identified the suspected violation as "deliberate accumulation of construction materials" and updated the processing status to "awaiting field evidence collection."

[0208] The second step involves field evidence gathering. The system generates standardized field verification tasks based on key geographical locations and risk indices. These tasks include verifying the presence of accumulated materials at location A50, their type and size, and whether they constitute illegal dumping. The tasks are simultaneously sent to the field personnel's mobile terminals, which automatically navigate to the target location. Upon arrival, the field personnel collect photos to cover the entire area of ​​the accumulated materials, details including the surrounding environment, and also capture video footage. They also record a text description stating, "Approximately 20 square meters of sand and gravel are piled up on site, without proper permits." This information is uploaded to the cultural relic safety monitoring platform via mobile terminals. Upon receiving the data, the system updates the processing status to "Pending legality determination."

[0209] The third step is to determine the legality. Based on the evidence collected in the field and the regulations for the protection of ancient city ruins, the management personnel determined that the pile-up was an illegal activity and updated the status to pending field verification and rectification.

[0210] The fourth step is field verification. After receiving the verification task, the field personnel will go to the site again to verify the cleanup of the accumulated materials and upload the information in real time. The status on the updated platform will be "Verification in progress". Once the rectification is confirmed, the status will be updated to "Restored". In the above process, the verification information needs to be uploaded in real time.

[0211] Figure 4 This is a schematic diagram illustrating a specific implementation of an integrated air-ground cultural relic security monitoring system provided in this application. (Refer to...) Figure 4 The system may include:

[0212] The acquisition module 41 is used to acquire radar image data, UAV image data and ground vibration data collected by the acquisition device deployed in the target site area at different time periods.

[0213] Module 42 is used to establish surface reference parameters for the target site area based on geological survey data and meteorological data of the target site area.

[0214] The identification module 43 is used to compare radar image data in different time periods to identify the surface change information of the target site area;

[0215] The first generation module 44 is used to generate a surface difference map of the target site area based on the degree of deviation between the surface change information and the surface reference parameters. The surface difference map is used to represent the deviation values ​​of each geographical location in the target site area.

[0216] The extraction module 45 is used to extract image slices corresponding to key geographical locations that meet preset deviation values ​​in the surface difference map from the UAV image data, and to perform multi-scale texture decomposition and color space conversion on the image slices to obtain the visual feature vectors of the image slices.

[0217] Analysis module 46 is used to combine the ground vibration data to perform spatiotemporal correlation analysis on the key geographical location and the visual feature vector, so as to calculate the risk index of the key geographical location;

[0218] The second generation module 47 is used to generate an alarm event if it is determined that there is illegal activity in the target site area based on the risk index.

[0219] This application provides an integrated air-ground cultural relic security monitoring system to implement the aforementioned integrated air-ground cultural relic security monitoring method. Therefore, the specific implementation of the integrated air-ground cultural relic security monitoring system can be found in the embodiment section of the integrated air-ground cultural relic security monitoring method above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0220] This application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described integrated air-ground cultural relic security monitoring methods.

[0221] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described integrated air-ground cultural relic security monitoring methods.

[0222] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0223] The embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above-described embodiments of the integrated air-ground cultural relic security monitoring method.

[0224] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0225] The above provides a detailed description of the integrated air-ground cultural relic security monitoring method and system provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for integrated air-ground monitoring of cultural relics, characterized in that, include: The system acquires radar imagery data and UAV imagery data of the target site area at different time periods, as well as ground vibration data collected by acquisition devices deployed within the target site area. Based on the geological survey data and meteorological data of the target site area, surface benchmark parameters for the target site area are established, including: The target archaeological site area is divided into multiple geographical locations; Based on the geological survey data, the geographical locations with the same geological attributes are classified into the corresponding geological zones; Historical meteorological data of the target site area were obtained, and the correlation between the historical meteorological data and the surface manifestations of various geological zones in historical image data was analyzed in order to construct a quantitative model reflecting the impact of meteorological changes on different geological zones. Based on real-time meteorological data at the time of image data acquisition for comparison, the corresponding benchmark value is calculated for each geological zone through the quantification model, and the benchmark value is assigned to all geographical locations within the geological zone to form a surface benchmark parameter covering each geographical location. By comparing radar image data from different time periods, surface change information of the target site area can be identified; Based on the degree of deviation between the surface change information and the surface reference parameters, a surface difference map of the target site area is generated, which is used to represent the deviation value of each geographical location in the target site area; For key geographical locations that meet the preset deviation value in the surface difference map, image slices corresponding to the key geographical locations are extracted from the UAV image data, and multi-scale texture decomposition and color space conversion are performed on the image slices to obtain the visual feature vectors of the image slices. By combining the ground vibration data, a spatiotemporal correlation analysis is performed on the key geographical location and the visual feature vector to calculate the risk index of the key geographical location; Based on the risk index, if it is determined that there are illegal activities within the target site area, an alarm event will be generated.

2. The method according to claim 1, characterized in that, The step of combining the ground seismic data to perform spatiotemporal correlation analysis on the key geographical location and the visual feature vector to calculate the risk index of the key geographical location includes: The occurrence time and spatial source of the vibration event are determined based on the ground vibration data; Using the occurrence time and spatial source of the vibration event as spatiotemporal constraints, a correlation analysis is performed on the key geographical locations in the surface difference map and the image slices corresponding to the visual feature vectors to calculate the risk index of the key geographical locations.

3. The method according to claim 1, characterized in that, The visual feature vectors include soil roughness, shadow sharpness, and deposit regularity; Perform multi-scale texture decomposition and color space transformation on the image slices to obtain the visual feature vectors of the image slices, including: The pixel color representation of the image slice is converted into a representation in which the luminance component and the color component are separated; The transformed image slices are decomposed into multi-level layers to separate the basic information layer representing the overall contour and multiple detail information layers representing different levels of refinement. Based on the multiple detailed information layers, soil roughness is calculated by analyzing the distribution intensity of high-frequency information; Based on the brightness components, the shadow sharpness is calculated by quantifying the rate of brightness change at the edges of the identified low-brightness areas. The regularity of the accumulation is calculated by analyzing the matching degree between the object contours identified in the image slices and the standard geometric shape, as well as the tortuosity of the contours themselves. The calculation results of soil roughness, shadow sharpness, and deposit regularity are combined to form a visual feature vector.

4. The method according to claim 2, characterized in that, The visual feature vectors include soil roughness, shadow sharpness, and deposit regularity; Using the occurrence time and spatial source point of the vibration event as spatiotemporal constraints, a correlation analysis is performed on key geographical locations in the surface difference map and the image slices corresponding to the visual feature vectors to calculate the risk index of the key geographical locations, including: Based on the occurrence time and spatial source point of the vibration event, determine the time window and spatial radius; Within the stated time window and spatial radius, key geographical locations are selected. A base risk score is determined based on the deviation value of each of the key geographical locations; The basic risk score is adjusted based on the visual feature vector associated with the key geographical location to obtain an adjusted risk score. The adjustment process includes increasing the basic risk score if the soil roughness, shadow sharpness, and deposit regularity meet preset conditions. The risk index is calculated based on the adjusted risk score and the vibration event level extracted from ground motion data.

5. The method according to claim 1, characterized in that, The comparison of radar image data within different time periods to identify surface change information in the target site area includes: Differential interferometry is used to process radar image data from different time periods to generate information on surface changes.

6. The method according to claim 1, characterized in that, Also includes: The key geographical location is marked on the digital map interface corresponding to the target site area, and the radar image data, the UAV image data and the risk index used for the determination are displayed in conjunction.

7. The method according to claim 6, characterized in that, Also includes: An online process for handling violations is initiated, which includes the steps of preliminary identification of activities at the key geographical locations, field evidence collection, legality determination, and field verification. Update the processing status of the violation based on the progress of the online processing procedure.

8. The method according to claim 7, characterized in that, Also includes: Based on the key geographical locations and risk indices, field verification tasks are generated and distributed to the mobile terminals of field personnel. Receive and record the verification results uploaded by the mobile terminal, which include on-site text, images, or video evidence; Update the status of the online processing procedure based on the on-site verification results.

9. A sky-ground integrated cultural relic safety monitoring system, characterized in that, include: The acquisition module is used to acquire radar image data, UAV image data and ground vibration data of the target site area at different time periods. The module is used to establish surface reference parameters for the target site area based on geological survey data and meteorological data, including: The target archaeological site area is divided into multiple geographical locations; Based on the geological survey data, the geographical locations with the same geological attributes are classified into the corresponding geological zones; Historical meteorological data of the target site area were obtained, and the correlation between the historical meteorological data and the surface manifestations of various geological zones in historical image data was analyzed in order to construct a quantitative model reflecting the impact of meteorological changes on different geological zones. Based on real-time meteorological data at the time of image data acquisition for comparison, the corresponding benchmark value is calculated for each geological zone through the quantification model, and the benchmark value is assigned to all geographical locations within the geological zone to form a surface benchmark parameter covering each geographical location. The identification module is used to compare radar image data within different time periods to identify surface change information in the target site area; The first generation module is used to generate a surface difference map of the target site area based on the degree of deviation between the surface change information and the surface reference parameters. The surface difference map is used to represent the deviation values ​​of each geographical location in the target site area. The extraction module is used to extract image slices corresponding to key geographical locations that meet preset deviation values ​​in the surface difference map from the UAV image data, and to perform multi-scale texture decomposition and color space conversion on the image slices to obtain the visual feature vectors of the image slices. The analysis module is used to combine the ground vibration data to perform spatiotemporal correlation analysis on the key geographical location and the visual feature vector, so as to calculate the risk index of the key geographical location. The second generation module is used to generate an alarm event if it is determined that there is illegal activity within the target site area based on the risk index.

Images