Magnetic target intelligent identification system and method based on perspective remote sensing technology

Abstract

The invention discloses a magnetic target intelligent identification system and method based on a perspective remote sensing technology, and the method comprises the following steps: extracting earth surface change information through optical satellite data to detect the earth surface change in an omnibearing manner, and delimiting a comprehensive target region according to the magnetic target information and the earth surface highlight abnormal information in SAR radar data; the unmanned aerial vehicle comprehensive detection platform is used for carrying out target area comprehensive detection on the limited comprehensive target area, and integrated detection of magnetic targets on the earth surface and the underground shallow layer and disguised magnetic targets are achieved; through cross validation of different methods and data, the problems of complex background interference, target camouflage and shielding, recognition accuracy and reliability under different weather and illumination conditions and the like are solved, integrated detection and discrimination of overground and underground shallow layers and camouflage magnetic objects are achieved, and technical support is provided for work such as illegal mining of mines.

Description

Technical Field

[0001] The present invention relates to the technical field of perspective remote sensing, and in particular to a magnetic target intelligent recognition system and method based on perspective remote sensing technology. Background Art

[0002] Remote sensing target recognition technology mainly relies on remote sensing technology, which is an emerging detection technology that can sense the properties and status of the target object from a distance without direct contact with the target object. It uses the characteristics of objects that can radiate, reflect or emit electromagnetic waves, and uses infrared, SAR / ISAR, hyperspectral / multispectral, lidar and other sensors to sense electromagnetic wave information from the target object from high altitude or long distance, and processes it into images or data through optical and electronic technologies to reveal the properties and status of the target.

[0003] Through remote sensing technology, we can obtain key information such as the specific location, type, quantity, etc. of the monitoring target, providing a decision-making basis for relevant departments' monitoring and law enforcement.

[0004] Specific to the technical methods, remote sensing target recognition and monitoring mainly include the following steps:

[0005] Data acquisition: remote sensing data of the target area is obtained through remote sensing platforms such as satellites and aircraft. These data may include various types of information such as visible light, infrared, and multi-spectral.

[0006] Data preprocessing: Preprocess the acquired remote sensing data, including radiation calibration, atmospheric correction, geometric correction and other steps to eliminate noise and distortion in the data and improve data quality.

[0007] Feature extraction: Extract feature information related to the target from the preprocessed data. These features may include shape, size, texture, spectral characteristics, etc. Feature extraction is a key step in target recognition, which directly affects the accuracy of subsequent target recognition.

[0008] Target recognition: Based on the extracted feature information, the target is identified using a classifier or deep learning algorithm. This step requires the design of appropriate algorithms and models to achieve accurate identification of different types of targets.

[0009] However, current remote sensing target recognition technology has technical problems such as complex background interference, target camouflage and occlusion, and unstable recognition under different weather and lighting conditions. Summary of the invention

[0010] The purpose of the present invention is to provide a magnetic target intelligent recognition system and method based on perspective remote sensing technology to solve the technical problems in the prior art of complex background interference, target camouflage and shielding, and unstable recognition under different weather and lighting conditions.

[0011] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:

[0012] A magnetic target intelligent recognition system based on perspective remote sensing technology, comprising:

[0013] Satellite detection system, used to monitor the area of ​​surface changes through optical satellites and SAR radar to delineate comprehensive target areas including newly added magnetic targets;

[0014] An unmanned aerial vehicle cruise system is integrated with a hyperspectral sensor and an unmanned aerial vehicle aeromagnetic sensor. The unmanned aerial vehicle cruise system performs cyclic monitoring in the comprehensive target area and generates monitoring data;

[0015] The data processing system is connected to the satellite detection system and the UAV cruise system in communication, and forms a perspective remote sensing data set based on the monitoring results. The perspective remote sensing data set is processed to delineate a comprehensive target area where magnetic targets exist, and to identify magnetic targets on the ground and in shallow underground layers within the comprehensive target area.

[0016] As a preferred solution of the present invention, the data processing system comprises:

[0017] A partition cutting module, used for performing two-dimensional modeling on the surface change area, and identifying a magnetic change area based on the surface change information of each partition of the surface change area;

[0018] A target recognition module, which forms a target sample set according to the detection results of the satellite detection system, trains the target sample set using a target recognition model, and identifies a comprehensive target area where a newly added magnetic target is located from the magnetic change area;

[0019] The magnetic target identification module, the UAV cruise system performs cyclic reciprocating monitoring on the comprehensive target area to form a monitoring data set, and identifies the surface magnetic targets and shallow-layer magnetic targets in the comprehensive target area by analyzing the monitoring data set.

[0020] In addition, the present invention also provides a magnetic target intelligent recognition method based on perspective remote sensing technology, comprising the following steps:

[0021] Step 100, using optical satellite data and SAR radar to extract surface change information to comprehensively detect surface changes in the surface change area, and based on the surface highlight abnormality information in the monitoring data of the optical satellite data and SAR radar, delineate a comprehensive target area based on the changes in the two-dimensional coordinate points of the surface highlight abnormality information;

[0022] Step 200: Use the drone cruise system to conduct comprehensive detection within the defined comprehensive target area to identify magnetic targets and camouflaged magnetic targets on the surface and in shallow underground layers.

[0023] As a preferred solution of the present invention, in step 100, the steps for defining the comprehensive target area are:

[0024] Step 101: Use the position information corresponding to the optical satellite to partition the surface change area, determine the two-dimensional coordinate range corresponding to each partition, and determine the target area where the magnetic change area is located based on the surface change information of the optical satellite data of each partition compared with the corresponding historical moment;

[0025] Step 102: Use SAR radar to detect the target area in sequence, divide the target area into different analysis areas based on the position information corresponding to each SAR radar, and determine the target reflectivity corresponding to each analysis area;

[0026] Step 103: Filter out areas with high-intensity reflection anomaly information from the SAR radar data of the analysis area, determine the two-dimensional coordinate range corresponding to the areas with high-intensity reflection anomaly information, compare the current areas with high-intensity reflection anomaly information with the historical areas with high-intensity reflection anomaly information, and use the area corresponding to the two-dimensional coordinate range corresponding to the area with the newly added high-intensity reflection anomaly information as the comprehensive target area.

[0027] As a preferred solution of the present invention, in step 101, a specific implementation method of identifying the target area where the magnetic change area is located is:

[0028] Pre-trained target recognition model: using the detection data of optical satellites to form historical optical remote sensing images, obtaining historical magnetic target data image images from the historical optical remote sensing images, and using the historical magnetic target data image images as the historical magnetic target sample set, using the Faster-R-CNN algorithm to train the historical magnetic target sample set to identify each newly generated magnetic target in the historical magnetic target sample set, and obtaining the final Faster-R-CNN model as the target recognition model;

[0029] Use the target recognition model to identify the target area in real time: Use the trained Faster-R-CNN model to detect the optical remote sensing image set acquired by the optical satellite in real time, output the magnetic targets corresponding to each optical remote sensing image set, determine the detection results of the newly added magnetic targets, and delineate the target area from the surface change area based on the detection results of the newly added magnetic targets.

[0030] As a preferred solution of the present invention, in step 103, the SAR radar performs magnetic target detection on the target area, and marks the two-dimensional coordinate points corresponding to the high-intensity reflection anomaly information in each analysis area;

[0031] Compare each two-dimensional coordinate point corresponding to high-intensity reflection anomaly information with the two-dimensional coordinate point corresponding to the historical high-intensity reflection anomaly information in the analysis area, and integrate the newly added two-dimensional coordinate point and the high-intensity reflection anomaly information corresponding to the two-dimensional coordinate point into the comprehensive target area;

[0032] The two-dimensional coordinate points with high-intensity reflection anomaly information in each analysis area are compared with the reflection data of the historical high-intensity reflection anomaly information of the two-dimensional coordinate points, and the two-dimensional coordinate points that are significantly stronger than the historical high-intensity reflection anomaly information are integrated into the comprehensive target area.

[0033] As a preferred solution of the present invention, in step 200, the method for using the drone cruise system to identify the magnetic targets on the surface and in the shallow underground layer in the comprehensive target area and the camouflaged magnetic targets is as follows:

[0034] Using the UAV cruise system to conduct comprehensive detection in the comprehensive target area, obtain the hyperspectral data and aeromagnetic data of the comprehensive target area, and cross-screen and judge the hyperspectral data and aeromagnetic data respectively to obtain the magnetic targets on the surface and shallow underground in the comprehensive target area, as well as the camouflaged magnetic targets in the comprehensive target area;

[0035] Processing the aeromagnetic data in the comprehensive target area to obtain high-intensity aeromagnetic anomaly information, marking them as suspected magnetic targets, and determining the two-dimensional coordinate point corresponding to each of the suspected magnetic targets;

[0036] Based on the spectral recognition technology, the surface information of the hyperspectral data in the comprehensive target area is classified to identify the material information of different surface objects, and the magnetic targets and camouflaged magnetic targets in the comprehensive target area are determined based on the material information, and the two-dimensional coordinate points of the magnetic targets are determined;

[0037] The two-dimensional coordinate points corresponding to the suspected magnetic targets identified based on the aeromagnetic data are matched with the two-dimensional coordinate points of the magnetic targets identified based on the spectral recognition technology, and the magnetic targets with completely matched coordinate values ​​form a magnetic target set.

[0038] As a preferred solution of the present invention, before processing the aeromagnetic data in the comprehensive target area, the aeromagnetic data is corrected according to the diurnal magnetic variation of the aeromagnetic data in the comprehensive target area to eliminate the interference of the solar diurnal magnetic field variation on the aeromagnetic measurement data.

[0039] As a preferred solution of the present invention, the method for correcting the aeromagnetic data according to the magnetic diurnal variation of the aeromagnetic data in the comprehensive target area is:

[0040] Obtaining diurnal variation data, which is obtained through observations at geomagnetic stations or seafloor geomagnetic diurnal variation observation stations;

[0041] Calculate the magnetic diurnal variation correction value. According to the acquired diurnal variation data, calculate the corresponding magnetic diurnal variation correction value, where the magnetic diurnal variation correction value is a fluctuation amount with a period of one day, which is used to describe the change of the geomagnetic field over time.

[0042] The aeromagnetic data are corrected point by point. The calculated magnetic diurnal variation correction value is applied to the aeromagnetic measurement data. Each measurement point is corrected point by point to eliminate the influence of magnetic diurnal variation on the aeromagnetic data.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention builds an unmanned aerial vehicle integrated detection system platform based on perspective remote sensing technology, uses a variety of detection methods to comprehensively analyze a variety of remote sensing data to identify magnetic targets in a comprehensive target area, thereby solving problems such as complex background interference, target camouflage and shielding, and recognition accuracy and reliability under different weather and lighting conditions through cross-validation of different methods and data, and realizing integrated recognition of magnetic targets on the ground and in shallow underground layers as well as camouflaged magnetic targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0046] Figure 1 A schematic diagram of the flow chart of the magnetic target intelligent recognition method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] like Figure 1 As shown, the present invention provides a magnetic target intelligent recognition system based on perspective remote sensing technology, comprising:

[0049] Satellite detection system, used to monitor the area of ​​surface changes through optical satellites and SAR radar to delineate comprehensive target areas including newly added magnetic targets;

[0050] The UAV cruise system integrates a hyperspectral sensor and a UAV aeromagnetic sensor. The UAV cruise system performs cyclic monitoring in the comprehensive target area and generates monitoring data.

[0051] The data processing system is connected to the satellite detection system and the UAV cruise system in communication, and forms a perspective remote sensing data set based on the monitoring results. By processing the perspective remote sensing data set, it can delineate the comprehensive target area where magnetic targets exist, and identify the magnetic targets on the ground and in the shallow underground layers within the comprehensive target area.

[0052] The data processing system includes:

[0053] A partition cutting module is used to perform two-dimensional modeling on the surface change area and identify the magnetic change area based on the surface change information of each partition of the surface change area;

[0054] The target recognition module forms a target sample set based on the detection results of the satellite detection system, trains the target sample set using the target recognition model, and identifies the comprehensive target area where the newly added magnetic targets are located from the magnetic change area;

[0055] The magnetic target recognition module and the UAV cruise system perform cyclical monitoring of the comprehensive target area to form a monitoring data set, and identify the surface magnetic targets and shallow-layer magnetic targets in the comprehensive target area by analyzing the monitoring data set.

[0056] This implementation method utilizes drones / satellites to integrate a series of remote sensing and geophysical exploration sensor equipment to form a drone / satellite comprehensive detection platform. The sensors include high-resolution optical satellites, SAR radars, hyperspectral sensors, and drone aeromagnetic sensors. The high-precision measurement data acquired by the platform provides data support for effective identification and authentication in complex environments.

[0057] Specifically, this embodiment uses optical satellite data to detect surface changes in all directions, and extracts surface change information. The detected change area is a large-scale area, and the surface change area caused by human activities is found. The target recognition model is then used to conduct further fine analysis of the surface change area caused by human activities to identify the newly added magnetic targets in the area where the surface changes caused by human activities are located, and the comprehensive target area with the newly added magnetic targets is delineated. The newly added magnetic targets in this embodiment are specifically large-scale machinery made of magnetic metal materials, thereby providing technical support for illegal mining and other work, wherein the range of surface changes must be larger than the range corresponding to the specified comprehensive target area.

[0058] The drone detection system is then used to identify magnetic targets in the integrated target area. Since the drone detection system is close to the integrated target area, the specific location of the magnetic target can be determined based on the identified object material information, and camouflaged magnetic targets can be identified, thereby providing assistance in guiding the specific location of large machinery made of magnetic metal materials.

[0059] Among them, the multispectral data in the satellite detection system is mainly based on high-resolution visible light data obtained by optical satellites, which is used to directly identify the target object. In this embodiment, the visible light data corresponding to the magnetic target is specifically determined, so as to determine whether there is a magnetic target based on the high-resolution visible light data obtained by the optical satellite.

[0060] The SAR radar data of the satellite detection system mainly uses SAR radar to obtain high-resolution SAR data, which is used to quickly extract suspected magnetic targets.

[0061] The hyperspectral data of the UAV detection system is mainly obtained by hyperspectral sensors with high-resolution hyperspectral data, which is used to identify the material information of different surface objects in order to identify magnetic targets.

[0062] The aeromagnetic data of the drone detection system uses aeromagnetic sensors to obtain aeromagnetic data, obtain surface magnetic anomalies, detect magnetic targets in shallow underground layers, and identify camouflaged magnetic targets.

[0063] This implementation method builds an unmanned aerial vehicle integrated detection system platform based on perspective remote sensing technology to achieve integrated detection of surface, shallow underground, and camouflaged magnetic targets; it achieves rapid detection and authentication of shallow magnetic targets on the ground and underground, improves recognition accuracy and reliability, reduces remote sensing satellite data information extraction operations, and improves the efficiency of magnetic target detection.

[0064] In addition, based on the above-mentioned magnetic target intelligent recognition system based on perspective remote sensing technology, the present invention also provides a magnetic target intelligent recognition method based on perspective remote sensing technology, comprising the following steps:

[0065] Step 100, using optical satellite data and SAR radar to extract surface change information to comprehensively detect surface changes in the surface change area, and based on the surface highlight anomaly information in the monitoring data of the optical satellite data and SAR radar, the comprehensive target area is delineated based on the changes in the two-dimensional coordinate points of the surface highlight anomaly information.

[0066] In step 100, the steps for defining the comprehensive target area are:

[0067] Step 101, use the position information corresponding to the optical satellite to partition the surface change area, determine the two-dimensional coordinate range corresponding to each partition, based on the surface change information that occurs compared with the optical satellite data of each partition and its corresponding historical moment, and determine the target area where the magnetic change area is located.

[0068] In step 101, the specific implementation method of identifying the target area where the magnetic change area is located is:

[0069] Pre-trained target recognition model: Use the detection data of optical satellites to form historical optical remote sensing images, and obtain historical magnetic target data images (excavators, transport vehicles, etc.) from the historical optical remote sensing images. The historical magnetic target data images in the historical optical remote sensing images are constantly added or reduced;

[0070] The historical magnetic target data images are used as the historical magnetic target sample set, and the Faster-R-CNN algorithm is used to train the historical magnetic target sample set to identify the newly generated magnetic targets or reduced magnetic targets in each historical magnetic target sample set. The most important thing is that the trained Faster-R-CNN model can identify the newly generated magnetic targets in the same detection area, so that the trained Faster-R-CNN model can identify the new excavators and transport vehicles corresponding to illegal mining in the mine, and the final Faster-R-CNN model is used as the target recognition model.

[0071] Use the target recognition model to identify the target area in real time: Use the trained Faster-R-CNN model to detect the optical remote sensing image set acquired by the optical satellite in real time, output the magnetic targets corresponding to each optical remote sensing image set, determine the detection results of the newly added magnetic targets, and delineate the target area from the surface change area based on the detection results of the newly added magnetic targets.

[0072] Step 102: Use SAR radar to detect the target area in sequence, divide the target area into different analysis areas based on the position information corresponding to each SAR radar, and determine the target reflectivity corresponding to each analysis area.

[0073] It should be noted that the reflectivity of the target measured by SAR radar is related to the physical properties of the target, such as the material, surface roughness and shape of the target. By analyzing the reflectivity of the target, the composition and material information of the target can be inferred. The high reflectivity of radar to metal objects is used.

[0074] Step 103: Filter out areas with high-intensity reflection anomaly information from the SAR radar data of the analysis area, determine the two-dimensional coordinate range corresponding to the areas with high-intensity reflection anomaly information, compare the current areas with high-intensity reflection anomaly information with the historical areas with high-intensity reflection anomaly information, and use the area corresponding to the two-dimensional coordinate range corresponding to the area with the newly added high-intensity reflection anomaly information as the comprehensive target area.

[0075] In step 103, the SAR radar performs magnetic target detection on the target area and marks the two-dimensional coordinate points corresponding to the high-intensity reflection anomaly information in each analysis area.

[0076] Compare each two-dimensional coordinate point corresponding to high-intensity reflection anomaly information with the two-dimensional coordinate point corresponding to the historical high-intensity reflection anomaly information in the analysis area, and integrate the newly added two-dimensional coordinate point and the high-intensity reflection anomaly information corresponding to the two-dimensional coordinate point into the comprehensive target area.

[0077] The two-dimensional coordinate points with high-intensity reflection anomaly information in each analysis area are compared with the reflection data of the historical high-intensity reflection anomaly information of the two-dimensional coordinate points, and the two-dimensional coordinate points that are significantly stronger than the historical high-intensity reflection anomaly information are integrated into the comprehensive target area.

[0078] Therefore, through the above processing method, the two-dimensional coordinate points corresponding to the newly added magnetic targets and the two-dimensional coordinate points corresponding to the old magnetic target points with obvious magnetic changes can be identified from the surface change area. The two-dimensional coordinate points corresponding to the above two types of changes are integrated into the comprehensive target area, and the comprehensive target area where the magnetic targets with magnetic changes are located can be accurately identified. The comprehensive target area is subjected to secondary comprehensive survey by the unmanned aerial vehicle cruise system to identify the magnetic targets and camouflaged magnetic targets on the surface and shallow underground of the comprehensive target area. Therefore, by combining the above-mentioned multiple magnetic target identification methods, the comprehensive target area with newly added magnetic targets and the two-dimensional coordinate points corresponding to each newly added magnetic target in the comprehensive target area can be effectively identified.

[0079] Step 200: Use the drone cruise system to conduct comprehensive detection within the defined comprehensive target area to identify magnetic targets and camouflaged magnetic targets on the surface and in shallow underground layers.

[0080] In step 200, the method for using the drone cruise system to identify the magnetic targets on the surface and in the shallow underground layer in the comprehensive target area and the camouflaged magnetic targets is as follows:

[0081] The drone cruise system is used to conduct comprehensive survey and detection within the comprehensive target area, obtain hyperspectral data and aeromagnetic data of the comprehensive target area, and cross-screen and judge the hyperspectral data and aeromagnetic data respectively to obtain magnetic targets on the surface and shallow underground in the comprehensive target area, as well as camouflaged magnetic targets in the comprehensive target area.

[0082] The aeromagnetic data in the comprehensive target area are processed to obtain high-intensity aeromagnetic anomaly information, which are marked as suspected magnetic targets, and the two-dimensional coordinate point corresponding to each suspected magnetic target is determined.

[0083] Based on spectral recognition technology, the surface information of the hyperspectral data in the comprehensive target area is classified, the material information of different surface objects is identified, and the magnetic targets and camouflaged magnetic targets in the comprehensive target area are determined based on the material information, and the two-dimensional coordinate points of the magnetic targets are determined.

[0084] Among them, the work of ground object classification using high-precision drone hyperspectral data is carried out. Based on hyperspectral remote sensing images and spectral recognition technology, the spectral characteristics of different materials in the wavelength range of 350-2500nm are used to classify surface information and extract target camouflage material information, providing support for target detection and authentication.

[0085] The two-dimensional coordinate points corresponding to the suspected magnetic targets identified based on the aeromagnetic data are matched with the two-dimensional coordinate points of the magnetic targets identified based on the spectral recognition technology, and the magnetic targets with completely matched coordinate values ​​form a magnetic target set.

[0086] Before processing the aeromagnetic data in the comprehensive target area, the aeromagnetic data are corrected according to the magnetic diurnal variation of the aeromagnetic data in the comprehensive target area to eliminate the interference of the solar diurnal magnetic field variation on the aeromagnetic measurement data.

[0087] The method for correcting the aeromagnetic data according to the magnetic daily variation of the aeromagnetic data in the comprehensive target area is as follows:

[0088] Obtaining diurnal variation data, which is obtained through observations at geomagnetic stations or seafloor geomagnetic diurnal variation observation stations;

[0089] Calculate the magnetic diurnal variation correction value. According to the acquired diurnal variation data, calculate the corresponding magnetic diurnal variation correction value, where the magnetic diurnal variation correction value is a fluctuation amount with a period of one day, which is used to describe the change of the geomagnetic field over time.

[0090] The aeromagnetic data are corrected point by point. The calculated magnetic diurnal variation correction value is applied to the aeromagnetic measurement data. Each measurement point is corrected point by point to eliminate the influence of magnetic diurnal variation on the aeromagnetic data.

[0091] In summary, the specific implementation method of this embodiment to effectively identify magnetic targets in a complex environment is as follows:

[0092] (1) First, we define the key work area from a large area, which is the area of ​​surface change;

[0093] (2) Identify target areas with magnetic targets from high-resolution optical satellite data of surface change areas using a target recognition model;

[0094] (3) Extract high-intensity reflection anomaly information from SAR satellite data to further limit the scope of the target area and form a comprehensive target area that includes new magnetic targets or original old magnetic targets with obvious magnetic changes;

[0095] (4) Use the UAV cruise system to conduct comprehensive survey and detection of the specified comprehensive target area to obtain the hyperspectral data and aeromagnetic data of the comprehensive target area;

[0096] (5) Obtain high-intensity aeromagnetic anomaly information of the comprehensive target area, calibrate it as a suspected magnetic target, classify the surface information based on the material information of the material based on the hyperspectral remote sensing image and spectral recognition technology of the comprehensive target area, and extract the magnetic targets and magnetic targets of the comprehensive target area;

[0097] (6) Match the two-dimensional coordinate points corresponding to the suspected magnetic targets identified based on the aeromagnetic data in step (5) above with the two-dimensional coordinate points of the magnetic targets identified based on the spectral recognition technology, and form a magnetic target set with the magnetic targets whose coordinate values ​​completely match.

[0098] The present invention is based on perspective remote sensing technology, integrates multi-spectral, hyperspectral, electromagnetic and other detection payloads on satellites and UAV airborne platforms, forms an integrated air-space detection platform, obtains perspective remote sensing data sets, and combines multiple detection methods to carry out identification and detection of magnetic targets. Through cross-validation of different methods and data, it solves problems such as complex background interference, target camouflage and obstruction, and identification accuracy and reliability under different weather and lighting conditions, realizes integrated detection and authentication of shallow and camouflaged magnetic targets on the ground, underground, and provides technical support for carrying out illegal mining and other work.

[0099] The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and protection scope of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present application.

Claims

1. A magnetic target intelligent recognition system based on perspective remote sensing technology, characterized in that: include: Satellite detection system, used to monitor the area of ​​surface changes through optical satellites and SAR radar to delineate comprehensive target areas including newly added magnetic targets; An unmanned aerial vehicle cruise system is integrated with a hyperspectral sensor and an unmanned aerial vehicle aeromagnetic sensor. The unmanned aerial vehicle cruise system performs cyclic monitoring in the comprehensive target area and generates monitoring data; The data processing system is connected to the satellite detection system and the UAV cruise system in communication, and forms a perspective remote sensing data set based on the monitoring results. The perspective remote sensing data set is processed to delineate a comprehensive target area where magnetic targets exist, and to identify magnetic targets on the ground and in shallow underground layers within the comprehensive target area.

2. According to claim 1, a magnetic target intelligent recognition system based on perspective remote sensing technology is characterized in that: The data processing system comprises: A partition cutting module, used for performing two-dimensional modeling on the surface change area, and identifying a magnetic change area based on the surface change information of each partition of the surface change area; A target recognition module, which forms a target sample set according to the detection results of the satellite detection system, trains the target sample set using a target recognition model, and identifies a comprehensive target area where a newly added magnetic target is located from the magnetic change area; The magnetic target identification module, the UAV cruise system performs cyclic reciprocating monitoring on the comprehensive target area to form a monitoring data set, and identifies the surface magnetic targets and shallow-layer magnetic targets in the comprehensive target area by analyzing the monitoring data set.

3. A magnetic target intelligent recognition method based on perspective remote sensing technology, characterized in that: A magnetic target intelligent recognition system based on perspective remote sensing technology according to any one of claims 1-2 comprises the following steps: Step 100, using optical satellite data and SAR radar to extract surface change information to comprehensively detect surface changes in the surface change area, and based on the surface highlight abnormality information in the monitoring data of the optical satellite data and SAR radar, delineate a comprehensive target area based on the changes in the two-dimensional coordinate points of the surface highlight abnormality information; Step 200: Use the drone cruise system to conduct comprehensive detection within the defined comprehensive target area to identify magnetic targets and camouflaged magnetic targets on the surface and in shallow underground layers.

4. The method for intelligently identifying magnetic targets based on perspective remote sensing technology according to claim 3 is characterized in that: In step 100, the steps for defining the comprehensive target area are: Step 101: Use the position information corresponding to the optical satellite to partition the surface change area, determine the two-dimensional coordinate range corresponding to each partition, and determine the target area where the magnetic change area is located based on the surface change information of the optical satellite data of each partition compared with the corresponding historical moment; Step 102: Use SAR radar to detect the target area in sequence, divide the target area into different analysis areas based on the position information corresponding to each SAR radar, and determine the target reflectivity corresponding to each analysis area; Step 103: Filter out areas with high-intensity reflection anomaly information from the SAR radar data of the analysis area, determine the two-dimensional coordinate range corresponding to the areas with high-intensity reflection anomaly information, compare the current areas with high-intensity reflection anomaly information with the historical areas with high-intensity reflection anomaly information, and use the area corresponding to the two-dimensional coordinate range corresponding to the area with the newly added high-intensity reflection anomaly information as the comprehensive target area.

5. The method for intelligently identifying magnetic targets based on perspective remote sensing technology according to claim 4 is characterized in that: In step 101, the specific implementation method of identifying the target area where the magnetic change area is located is: Pre-trained target recognition model: using the detection data of optical satellites to form historical optical remote sensing images, obtaining historical magnetic target data image images from the historical optical remote sensing images, and using the historical magnetic target data image images as the historical magnetic target sample set, using the Faster-R-CNN algorithm to train the historical magnetic target sample set to identify each newly generated magnetic target in the historical magnetic target sample set, and obtaining the final Faster-R-CNN model as the target recognition model; Use the target recognition model to identify the target area in real time: Use the trained Faster-R-CNN model to detect the optical remote sensing image set acquired by the optical satellite in real time, output the magnetic targets corresponding to each optical remote sensing image set, determine the detection results of the newly added magnetic targets, and delineate the target area from the surface change area based on the detection results of the newly added magnetic targets.

6. The method for intelligently identifying magnetic targets based on perspective remote sensing technology according to claim 5 is characterized in that: In step 103, the SAR radar performs magnetic target detection on the target area, and marks the two-dimensional coordinate points corresponding to the high-intensity reflection anomaly information in each analysis area; Compare each two-dimensional coordinate point corresponding to high-intensity reflection anomaly information with the two-dimensional coordinate point corresponding to the historical high-intensity reflection anomaly information in the analysis area, and integrate the newly added two-dimensional coordinate point and the high-intensity reflection anomaly information corresponding to the two-dimensional coordinate point into the comprehensive target area; The two-dimensional coordinate points with high-intensity reflection anomaly information in each analysis area are compared with the reflection data of the historical high-intensity reflection anomaly information of the two-dimensional coordinate points, and the two-dimensional coordinate points that are significantly stronger than the historical high-intensity reflection anomaly information are integrated into the comprehensive target area.

7. The method for intelligently identifying magnetic targets based on perspective remote sensing technology according to claim 6 is characterized in that: In step 200, the method for using the drone cruise system to identify the magnetic targets on the surface and in the shallow underground layer in the comprehensive target area and the camouflaged magnetic targets is as follows: Using the UAV cruise system to conduct comprehensive detection in the comprehensive target area, obtain the hyperspectral data and aeromagnetic data of the comprehensive target area, and cross-screen and judge the hyperspectral data and aeromagnetic data respectively to obtain the magnetic targets on the surface and shallow underground in the comprehensive target area, as well as the camouflaged magnetic targets in the comprehensive target area; Processing the aeromagnetic data in the comprehensive target area to obtain high-intensity aeromagnetic anomaly information, marking them as suspected magnetic targets, and determining the two-dimensional coordinate point corresponding to each of the suspected magnetic targets; Based on the spectral recognition technology, the surface information of the hyperspectral data in the comprehensive target area is classified to identify the material information of different surface objects, and the magnetic targets and camouflaged magnetic targets in the comprehensive target area are determined based on the material information, and the two-dimensional coordinate points of the magnetic targets are determined; The two-dimensional coordinate points corresponding to the suspected magnetic targets identified based on the aeromagnetic data are matched with the two-dimensional coordinate points of the magnetic targets identified based on the spectral recognition technology, and the magnetic targets with completely matched coordinate values ​​form a magnetic target set.

8. The method for intelligently identifying magnetic targets based on perspective remote sensing technology according to claim 7 is characterized in that: Before processing the aeromagnetic data in the comprehensive target area, the aeromagnetic data is corrected according to the diurnal magnetic variation of the aeromagnetic data in the comprehensive target area to eliminate the interference of the solar diurnal magnetic field variation on the aeromagnetic measurement data.

9. The method for intelligently identifying magnetic targets based on perspective remote sensing technology according to claim 8, characterized in that: The method for correcting the aeromagnetic data according to the magnetic diurnal variation of the aeromagnetic data in the comprehensive target area is as follows: Obtaining diurnal variation data, which is obtained through observations at geomagnetic stations or seafloor geomagnetic diurnal variation observation stations; Calculate the magnetic diurnal variation correction value. According to the acquired diurnal variation data, calculate the corresponding magnetic diurnal variation correction value, where the magnetic diurnal variation correction value is a fluctuation amount with a period of one day, which is used to describe the change of the geomagnetic field over time. The aeromagnetic data are corrected point by point. The calculated magnetic diurnal variation correction value is applied to the aeromagnetic measurement data. Each measurement point is corrected point by point to eliminate the influence of magnetic diurnal variation on the aeromagnetic data.

Images