Mine ground object accurate identification method based on multi-source remote sensing data fusion technology

Abstract

The invention discloses a mine surface feature accurate identification method based on a multi-source remote sensing data fusion technology, and relates to the technical field of data identification, and the method comprises the following implementation steps: configuring the IP address information of a remote control area server of a mine surface feature; entering a data acquisition end, and collecting time sequence multi-source remote sensing data from different sensors and platforms in real time; entering a data fusion end, and eliminating the influence of sensor errors on data quality in real time; and 4, entering a comprehensive identification end, extracting different ground feature features in real time, and verifying and optimizing a result in real time after comprehensive identification. According to the mine ground feature accurate identification method based on the multi-source remote sensing data fusion technology, remote sensing data early warning signals are sent in time, the accuracy of mine ground feature information acquisition is guaranteed, disaster early warning signals are sent to mine ground features, the safety and reliability of mine ground feature identification are guaranteed, accurate identification of strip mine mountain features is achieved, and the accuracy of mine ground feature identification is improved. And powerful support is provided for mine management and decision making.

Description

Technical Field

[0001] The present invention relates to the technical field of data recognition, and in particular to a method for accurately identifying mining objects based on multi-source remote sensing data fusion technology. Background Art

[0002] Mine feature identification is a technology that uses remote sensing technology, machine learning, and hyperspectral analysis to accurately identify geological features and mineral resources in a mining area. Mine feature identification refers to the use of various technical means, such as remote sensing images and hyperspectral imaging, to identify and distinguish different surface features in a mining area. This process is of great significance for the exploration, development, and management of mineral resources. With the development of remote sensing technology and computer vision technology, especially the application of deep learning, the accuracy and efficiency of mine feature identification have been significantly improved.

[0003] Publication No. CN202080000930.0 discloses a method and platform for dynamic monitoring of typical landforms in mining based on multi-source remote sensing data fusion and deep neural network. This method can improve the spatial resolution of remote sensing data images and enhance the images to make it easier to identify typical landforms. At the same time, the machine interpretation method used can efficiently and accurately identify typical landforms, so that typical landforms in mines can be monitored online in real time.

[0004] After searching the above patents, it was found that although the typical mine features can be monitored online in real time during mine feature identification, the online monitoring values ​​of the mine features cannot provide early warning of the changing trends of the mine features due to the constantly changing environment. It is impossible to ensure the accurate identification of mine features, and it is also difficult to provide early warning of mine features based on the online monitoring data.

[0005] Therefore, a new method for accurate identification of mining objects based on multi-source remote sensing data fusion technology is proposed to solve the above problems. Summary of the invention

[0006] The main purpose of the present invention is to provide a method for accurately identifying mine features based on multi-source remote sensing data fusion technology to solve the problems raised in the above background.

[0007] To achieve the above object, the technical solution adopted by the present invention is: a method for accurately identifying mining objects based on multi-source remote sensing data fusion technology, the method comprising the following implementation steps:

[0008] Step 1: Configure the IP address information of the mine feature remote control area server;

[0009] Step 2: Enter the data collection terminal, collect time-series multi-source remote sensing data from different sensors and platforms in real time, and analyze the collected remote sensing data in real time to determine whether the remote sensing data is abnormal and whether the data quality is abnormal in real time;

[0010] Step 3: Enter the data fusion end, correct the remote sensing data with abnormal data quality in real time, eliminate the impact of sensor errors on data quality in real time, and use image fusion to organically integrate multi-source heterogeneous remote sensing data, build a deep neural network model, and extract the features of ground objects in the image in real time;

[0011] Step 4: Enter the comprehensive recognition terminal, divide the mining area into different types of land objects through image segmentation technology, extract the characteristics of different land objects in real time, and verify and optimize the results in real time after comprehensive recognition.

[0012] The data collection terminal includes a data collection module, a data analysis module and a data tracking module;

[0013] The data acquisition module includes a ground measured data unit and an auxiliary data unit;

[0014] The ground measured data unit is used to collect the current optical image, radar image, infrared image and ground measured data in real time through satellite remote sensing, UAV remote sensing and ground sensor data sources, and record them in real time through a data recorder;

[0015] The auxiliary data unit is used to collect basic geographic information, mining plans and historical monitoring records of the mining area in real time through a data receiver and a data recorder.

[0016] The data analysis module includes a data acquisition unit and an abnormality determination unit;

[0017] The data acquisition unit is used to receive the basic geographic information, mining plan, historical monitoring records and current monitoring records of the mining area in real time through the data receiver;

[0018] The abnormality judgment unit is used to use the calculation formula to judge in real time whether the data monitored and recorded at the current moment needs to be corrected by the remote sensing data at the current moment. The monitoring record at the current moment includes the atmospheric factor correction value, radiation correction value and geometric correction value of the mine feature. First, the vegetation index calculation formula considering the atmospheric influence of the enhanced mine feature vegetation index 2 is used, combined with the normalized difference vegetation index and the atmospheric correction factor ATB to reduce the interference of the atmosphere on the vegetation monitoring in the mine feature. The formula is: enhanced mine feature vegetation index 2 = 2.5*(NIR-RED) / (NIR+2.4*RED+1)*(1-ATB);

[0019] Among them, NIR is the near-infrared band reflectance, RED is the red band reflectance, and ATB is the atmospheric correction factor;

[0020] The calculation formula for determining the abnormal data of the monitoring record at the current moment is as follows:

[0021]

[0022] Among them, P n Indicates the judgment data value, U 1 Indicates the data value before the current monitoring record correction, U 2 represents the corrected data value of the current monitoring record, x represents the standard remote sensing data energy consumption of the data value, and X represents the standard remote sensing data energy consumption of the corrected data value. If the calculated E n = 0, the data does not need to be corrected. If E n ≠0, it is determined that the data needs to be corrected.

[0023] The data tracking module includes a data tracking unit and an abnormal warning unit;

[0024] The data tracking unit is used to track the monitoring record data value at the current moment in real time through a data tracker;

[0025] The abnormal warning unit is used to send a voice alarm to the reporting system when it is determined that the data needs to be corrected, and to record the abnormal data that needs to be corrected in real time through color identification.

[0026] The data fusion end includes a data processing module, a data fusion module and a model building module;

[0027] The data processing module includes a data correction unit, which is used to correct the heterogeneous data that is determined to need correction. The correction formula is as follows:

[0028]

[0029] Among them, U, V and x, y are the image coordinates of the data value before and after the transformation, P ij and Q ij are the polynomial correction coefficients, n represents the polynomial degree, and the calculated x and y represent the geometric correction data values.

[0030] The data fusion module includes an image generation unit and an image extraction unit;

[0031] The image generating unit is used to automatically generate remote sensing images through image fusion technology;

[0032] The image extraction unit is used to determine what the mine feature area on the remote sensing image is based on the mine feature shape features on the remote sensing image. The determination steps are as follows:

[0033] According to the first point of the mine feature shape on the remote sensing image, the equidistant feature points of the same shape of the mine feature and the currently tracked feature points of the mine feature shape, the average distance of the equidistant feature points of the mine feature shape is calculated. The currently tracked feature points are the feature point positions of the remote sensing image in different time periods. The average distance of the feature points is the distance between the mine feature in different time periods and the standard mine feature area of ​​the remote sensing image. Then, according to the different point values ​​of the feature points at different times, the approximate outline of the mine feature in the current remote sensing image is calculated, and the average distance Z of the matching feature points is calculated. t The calculation formula is as follows:

[0034]

[0035] Among them, f x,y-1 Represents the characteristic points of the shape of the mine features in the specified area of ​​the remote sensing image at time t-1, f x,y represents the feature point monitored and recorded at time t, M represents the distance between the feature points tracked at time t, x represents the value of a point in the X-axis direction, and y represents the value of the moving position in the Y-axis direction;

[0036] By moving the pixel point on the known path at time t-1 by Z t , the approximate outline of the objects in the mine features of the remote sensing image at time t can be obtained.

[0037] The model building module includes a model building unit, a manual interpretation unit, an online monitoring unit and a trend warning unit;

[0038] The model building unit is used to construct a convolutional neural network model, which is used to set mine feature identification parameters based on multi-source remote sensing data fusion technology, and the mine feature identification parameters include standard values ​​of environmental parameters, disaster warning parameter levels, mineral distribution range parameters, vegetation destruction parameters and soil and water loss parameters;

[0039] The manual interpretation unit is used to introduce manual visual interpretation as an auxiliary means to verify and correct the results of machine interpretation;

[0040] The online monitoring unit is used to monitor the basic parameters of the current mining area in real time through data monitoring equipment. The basic parameters include environmental parameters, disaster level, mineral distribution range, vegetation destruction rate, soil erosion rate and historical parameter difference;

[0041] The trend warning unit is used to create, manage, analyze and draw each remote sensing image area through GIS, and predict the development trend of mining objects in real time through the mining object monitoring parameters in the remote sensing image area. The prediction method is as follows:

[0042] Step I: Set the planning standard parameters for the mine feature planning project in the corresponding remote sensing image area, use earthquake monitoring equipment, landslide monitoring equipment, debris flow monitoring equipment, ground subsidence monitoring equipment and volcanic eruption monitoring equipment, temperature sensors, humidity sensors and light sensors to detect the execution parameters of the corresponding mine feature planning area in real time, and analyze the environmental parameters, disaster level, mineral distribution range, vegetation destruction rate, soil erosion rate and historical parameter difference of the corresponding mine feature planning area in real time;

[0043] Step II: Calculate the difference between the execution parameters of the corresponding area and the planned standard parameters. If the difference = ±0.002, it is judged that the planned parameters tend to the safe value. If the difference > 0.002 or < -0.002, it is judged that the planned parameters exceed the safe value, and a planned trend warning signal is issued. Seven days is a cycle. The development trend of the mining features in the corresponding area is predicted based on the planned parameters of three cycles. If the average values ​​of the planned parameters of the three cycles do not exceed the planned standard parameters, the mining feature development plan of the corresponding area is judged to be safe and stable. If the average values ​​of the planned parameters of the three cycles exceed the planned standard parameters, the mining feature development plan of the corresponding area is judged to be abnormal.

[0044] The comprehensive recognition end includes an image segmentation module, a feature extraction module, a feature classification module and a comprehensive recognition module;

[0045] The image segmentation module includes a land feature type unit, which is used to divide the mine land feature area into different land feature types by calculating the approximate contour. The land feature types include mine pits, waste rock piles and tailings ponds.

[0046] The feature extraction module is used to extract the spectrum, texture and shape features in the mining area through the calculation formula of the image extraction unit;

[0047] The feature classification module is used to classify the extracted mine features into different categories, including mine pits, waste rock piles, rivers, buildings and tailings ponds.

[0048] The comprehensive analysis module includes a multi-source data combination unit and a data verification unit;

[0049] The multi-source data combination unit is used to combine remote sensing data with GIS, use the spatial analysis function of GIS to perform spatial distribution and statistical analysis on mining objects, and combine remote sensing data with ground survey data to perform difference calculation between the data analysis value at the current moment and the trend warning value in real time. If the difference is 0, the data analysis is judged to be valid, and if the difference is ≠ 0, the data analysis is judged to be invalid, and data correction is performed twice;

[0050] The data verification unit is used to perform data verification in real time on the data results of the multi-source data structure unit through a data anomaly calculation formula to correct the data combination and data recognition results.

[0051] The present invention has the following beneficial effects:

[0052] 1. In the present invention, by setting a data acquisition terminal, when accurately identifying mining features, the collected remote sensing data is analyzed in real time, and the remote sensing data collected at the current moment is used in real time to determine whether the remote sensing data is abnormal, and a remote sensing data early warning signal is issued in time to ensure the accuracy of mining feature information acquisition and reduce data errors in the process of obtaining mining feature information;

[0053] 2. In the present invention, by setting up a data fusion terminal, when accurately identifying mining objects, remote sensing data with abnormal data quality is corrected in real time, and multi-source heterogeneous remote sensing data is organically integrated by image fusion, and the features of objects in remote sensing images are extracted in real time. Trends of monitoring data are predicted in real time, and disaster warning signals are issued to mining objects in advance to ensure the safety and reliability of mining object identification;

[0054] 3. In the present invention, by setting up a comprehensive identification terminal, for accurate identification of mine features, the spatial analysis function of GIS is used to perform spatial distribution and statistical analysis of mine features, and remote sensing data is combined with ground survey data to extract different features of features in real time. After comprehensive identification, the results are verified and optimized in real time to achieve accurate identification of open-pit mine features, providing strong support for mine management and decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is an overall flow chart of the method for accurate identification of mining objects based on multi-source remote sensing data fusion technology of the present invention;

[0056] Figure 2 It is a schematic diagram of the architecture of the data acquisition end of the method for accurate identification of mining objects based on multi-source remote sensing data fusion technology of the present invention;

[0057] Figure 3 It is a schematic diagram of the architecture of the data fusion end of the method for accurate identification of mining objects based on multi-source remote sensing data fusion technology of the present invention;

[0058] Figure 4 It is a schematic diagram of the architecture of the comprehensive identification end of the method for accurate identification of mining objects based on multi-source remote sensing data fusion technology of the present invention. DETAILED DESCRIPTION

[0059] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0060] Embodiment 1

[0061] Please refer to Figure 1 to Figure 2 As shown in the figure: The method for accurate identification of mining objects based on multi-source remote sensing data fusion technology includes the following implementation steps:

[0062] Step 1: Configure the IP address information of the mine feature remote control area server;

[0063] Step 2: Enter the data collection terminal, collect time-series multi-source remote sensing data from different sensors and platforms in real time, and analyze the collected remote sensing data in real time to determine whether the remote sensing data is abnormal and whether the data quality is abnormal in real time;

[0064] Step 3: Enter the data fusion end, correct the remote sensing data with abnormal data quality in real time, eliminate the impact of sensor errors on data quality in real time, and use image fusion to organically integrate multi-source heterogeneous remote sensing data, build a deep neural network model, and extract the features of ground objects in the image in real time;

[0065] Step 4: Enter the comprehensive recognition terminal, divide the mining area into different types of land objects through image segmentation technology, extract the characteristics of different land objects in real time, and verify and optimize the results in real time after comprehensive recognition.

[0066] The data collection end includes a data collection module, a data analysis module and a data tracking module;

[0067] The data acquisition module includes a ground measured data unit and an auxiliary data unit;

[0068] The ground measured data unit is used to collect the current optical image, radar image, infrared image and ground measured data in real time through satellite remote sensing, UAV remote sensing and ground sensor data sources, and record them in real time through a data recorder;

[0069] The auxiliary data unit is used to collect basic geographic information, mining plans and historical monitoring records of the mine area in real time through data receivers and data recorders.

[0070] The data analysis module includes a data acquisition unit and an abnormality determination unit;

[0071] The data acquisition unit is used to receive the basic geographic information, mining plan, historical monitoring records and current monitoring records of the mining area in real time through the data receiver;

[0072] The abnormality judgment unit is used to use the calculation formula to judge in real time whether the data monitored and recorded at the current moment needs to be corrected. The monitoring record at the current moment includes the atmospheric factor correction value, radiation correction value and geometric correction value of the mine feature. First, the vegetation index calculation formula considering the atmospheric influence of the enhanced mine feature vegetation index 2 is used, combined with the normalized difference vegetation index and the atmospheric correction factor ATB to reduce the interference of the atmosphere on the vegetation monitoring in the mine feature. The formula is: enhanced mine feature vegetation index 2 = 2.5*(NIR–RED) / (NIR+2.4*RED+1)*(1-ATB);

[0073] Among them, NIR is the near-infrared band reflectance, RED is the red band reflectance, and ATB is the atmospheric correction factor;

[0074] The calculation formula for determining the abnormal data of the monitoring record at the current moment is as follows:

[0075]

[0076] Among them, P n Indicates the judgment data value, U 1 Indicates the data value before the current monitoring record correction, U 2 represents the corrected data value of the current monitoring record, x represents the standard remote sensing data energy consumption of the data value, and X represents the standard remote sensing data energy consumption of the corrected data value. If the calculated E n = 0, the data does not need to be corrected. If E n ≠0, it is determined that the data needs to be corrected.

[0077] The data tracking module includes a data tracking unit and an abnormal warning unit;

[0078] The data tracking unit is used to track the monitoring and recording data value at the current moment in real time through the data tracker;

[0079] The abnormal warning unit is used to issue a voice alarm to the reporting system when it determines that the data needs to be corrected, and to record the abnormal data that needs to be corrected in real time through color identification.

[0080] By collecting time-series multi-source remote sensing data from different sensors and platforms in real time, and analyzing the collected remote sensing data in real time, it is possible to determine whether the remote sensing data is abnormal through the remote sensing data collected at the current moment, and to track the current data of the remote sensing data in real time, and to issue remote sensing data early warning signals in a timely manner, the accuracy of mine feature information acquisition is guaranteed, and data errors in the process of obtaining mine feature information are reduced.

[0081] Embodiment 2

[0082] Please refer to Figure 3As shown: Based on the first embodiment, the data fusion end includes a data processing module, a data fusion module and a model building module;

[0083] The data processing module includes a data correction unit, which is used to correct the heterogeneous data that needs to be corrected. The correction formula is as follows:

[0084]

[0085] Among them, U, V and x, y are the image coordinates of the data value before and after the transformation, P ij and Q ij are the polynomial correction coefficients, n represents the polynomial degree, and the calculated x and y represent the geometric correction data values.

[0086] The data fusion module includes an image generation unit and an image extraction unit;

[0087] The image generation unit is used to automatically generate remote sensing images through image fusion technology;

[0088] The image extraction unit is used to determine what the mine feature area on the remote sensing image is based on the mine feature shape features on the remote sensing image. The determination steps are as follows:

[0089] According to the first point of the mine feature shape on the remote sensing image, the equidistant feature points of the same shape of the mine feature and the currently tracked feature points of the mine feature shape, the average distance of the equidistant feature points of the mine feature shape is calculated. The currently tracked feature points are the feature point positions of the remote sensing image in different time periods. The average distance of the feature points is the distance between the mine feature in different time periods and the standard mine feature area of ​​the remote sensing image. Then, according to the different point values ​​of the feature points at different times, the approximate outline of the mine feature in the current remote sensing image is calculated, and the average distance Z of the matching feature points is calculated. t The calculation formula is as follows:

[0090]

[0091] Among them, f x,y-1 Represents the characteristic points of the shape of the mine features in the specified area of ​​the remote sensing image at time t-1, f x,y represents the feature point monitored and recorded at time t, M represents the distance between the feature points tracked at time t, x represents the value of a point in the X-axis direction, and y represents the value of the moving position in the Y-axis direction;

[0092] By moving the pixel point on the known path at time t-1 by Z t , the approximate outline of the objects in the mine features of the remote sensing image at time t can be obtained.

[0093] The model building module includes model building unit, manual interpretation unit, online monitoring unit and trend warning unit;

[0094] The model building unit is used to construct a convolutional neural network model, which is used to set the mine feature identification parameters based on the multi-source remote sensing data fusion technology. The mine feature identification parameters include standard values ​​of environmental parameters, disaster warning parameter levels, mineral distribution range parameters, vegetation destruction parameters and soil and water loss parameters;

[0095] The manual interpretation unit is used to introduce manual visual interpretation as an auxiliary means to verify and correct the results of machine interpretation;

[0096] The online monitoring unit is used to monitor the basic parameters of the current mining area in real time through data monitoring equipment. The basic parameters include environmental parameters, disaster level, mineral distribution range, vegetation destruction rate, soil erosion rate and historical parameter difference;

[0097] The trend warning unit is used to create, manage, analyze and draw each remote sensing image area through GIS, and predict the development trend of mining features in real time through the mining feature monitoring parameters in the remote sensing image area. The prediction method is as follows:

[0098] Step I: Set the planning standard parameters for the mine feature planning project in the corresponding remote sensing image area, use earthquake monitoring equipment, landslide monitoring equipment, debris flow monitoring equipment, ground subsidence monitoring equipment and volcanic eruption monitoring equipment, temperature sensors, humidity sensors and light sensors to detect the execution parameters of the corresponding mine feature planning area in real time, and analyze the environmental parameters, disaster level, mineral distribution range, vegetation destruction rate, soil erosion rate and historical parameter difference of the corresponding mine feature planning area in real time;

[0099] Step II: Calculate the difference between the execution parameters of the corresponding area and the planned standard parameters. If the difference = ±0.002, it is judged that the planned parameters tend to the safe value. If the difference > 0.002 or < -0.002, it is judged that the planned parameters exceed the safe value, and a planned trend warning signal is issued. Seven days is a cycle. The development trend of the mining features in the corresponding area is predicted based on the planned parameters of three cycles. If the average values ​​of the planned parameters of the three cycles do not exceed the planned standard parameters, the mining feature development plan of the corresponding area is judged to be safe and stable. If the average values ​​of the planned parameters of the three cycles exceed the planned standard parameters, the mining feature development plan of the corresponding area is judged to be abnormal.

[0100] By collecting time-series multi-source remote sensing data from different sensors and platforms in real time, and analyzing the collected remote sensing data in real time, it is possible to determine whether the remote sensing data is abnormal through the remote sensing data collected at the current moment, and to track the current data of the remote sensing data in real time, and to issue remote sensing data early warning signals in a timely manner, the accuracy of mine feature information acquisition is guaranteed, and data errors in the process of obtaining mine feature information are reduced.

[0101] Embodiment 3

[0102] Please refer to Figure 4 As shown: Based on the first embodiment, the comprehensive recognition end includes an image segmentation module, a feature extraction module, a feature classification module and a comprehensive recognition module;

[0103] The image segmentation module includes a land feature type unit, which is used to divide the mine land feature area into different land feature types by calculating the approximate contour. The land feature types include mine pits, waste rock piles and tailings ponds.

[0104] The feature extraction module is used to extract the spectrum, texture and shape features in the mining area through the calculation formula of the image extraction unit;

[0105] The feature classification module is used to classify the extracted mining features into different categories, including mine pits, waste rock piles, rivers, buildings and tailings ponds.

[0106] The comprehensive analysis module includes a multi-source data combination unit and a data verification unit;

[0107] The multi-source data combination unit is used to combine remote sensing data with GIS, and use the spatial analysis function of GIS to perform spatial distribution and statistical analysis on mining objects. At the same time, remote sensing data is combined with ground survey data to calculate the difference between the current data analysis value and the trend warning value in real time. If the difference = 0, the data analysis is judged to be valid. If the difference ≠ 0, the data analysis is judged to be invalid, and data correction is performed twice;

[0108] The data verification unit is used to perform real-time data verification on the data results of the multi-source data structure unit through a data anomaly calculation formula to correct the data combination and data recognition results.

[0109] By extracting the characteristics of different land objects in real time, comprehensively identifying them and verifying and optimizing the results in real time, using the spatial analysis function of GIS to conduct spatial distribution and statistical analysis of mine objects, and through image segmentation, feature extraction and classification extraction recognition, as well as combining remote sensing data with ground survey data, the characteristics of different land objects are extracted in real time, comprehensively identifying them and verifying and optimizing the results in real time, accurate identification of open-pit mine objects is achieved, providing strong support for mine management and decision-making.

[0110] In the present invention, a method for accurately identifying mine features based on multi-source remote sensing data fusion technology first configures the IP address information of the remote control area server of the mine features; enters the data acquisition end, collects time-series multi-source remote sensing data from different sensors and platforms in real time, and analyzes the collected remote sensing data in real time to determine whether the remote sensing data is abnormal, and judges whether the data quality is abnormal in real time, collects time-series multi-source remote sensing data from different sensors and platforms in real time, and analyzes the collected remote sensing data in real time, and determines whether the remote sensing data is abnormal through the remote sensing data collected at the current moment in real time, and tracks the current data of the remote sensing data in real time, and issues remote sensing data early warning signals in time to ensure the accuracy of mine feature information acquisition and reduce data errors in the process of mine feature information acquisition; enters the data fusion end, corrects remote sensing data with abnormal data quality in real time, eliminates the influence of sensor errors on data quality in real time, and uses image fusion to organically fuse multi-source heterogeneous remote sensing data, constructs a deep neural network model, extracts feature of features in the image in real time, and Remote sensing data with abnormal quality are corrected in real time to promptly eliminate the impact of sensor errors on data quality. Image fusion is used to organically integrate multi-source heterogeneous remote sensing data, build a deep neural network model, extract the features of objects in remote sensing images in real time, and introduce artificial visual interpretation to verify and correct the results of machine interpretation in real time. Online dynamic monitoring of open-pit mine objects is carried out, and trend prediction of monitoring data is carried out in real time. Disaster warning signals are issued to mine objects in advance to ensure the safety and reliability of mine object identification. At the comprehensive identification end, the mine area is divided into different types of objects through image segmentation technology, and the features of different objects are extracted in real time. After comprehensive identification, the results are verified and optimized in real time. The spatial distribution and statistical analysis of mine objects are carried out by using the spatial analysis function of GIS. Image segmentation, feature extraction and classification extraction recognition, as well as the combination of remote sensing data and ground survey data, extract the features of different objects in real time, and verify and optimize the results in real time after comprehensive identification, so as to realize the accurate identification of open-pit mine objects and provide strong support for mine management and decision-making.

[0111] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for accurately identifying mining objects based on multi-source remote sensing data fusion technology, characterized in that: The method comprises the following implementation steps: Step 1: Configure the IP address information of the mine feature remote control area server; Step 2: Enter the data collection terminal, collect time-series multi-source remote sensing data from different sensors and platforms in real time, and analyze the collected remote sensing data in real time to determine whether the remote sensing data is abnormal and whether the data quality is abnormal in real time; Step 3: Enter the data fusion end, correct the remote sensing data with abnormal data quality in real time, eliminate the impact of sensor errors on data quality in real time, and use image fusion to organically integrate multi-source heterogeneous remote sensing data, build a deep neural network model, and extract the features of ground objects in the image in real time; Step 4: Enter the comprehensive recognition terminal, divide the mining area into different types of land objects through image segmentation technology, extract the characteristics of different land objects in real time, and verify and optimize the results in real time after comprehensive recognition.

2. The method according to claim 1, characterized in that: The data collection terminal includes a data collection module, a data analysis module and a data tracking module; The data acquisition module includes a ground measured data unit and an auxiliary data unit; The ground measured data unit is used to collect the current optical image, radar image, infrared image and ground measured data in real time through satellite remote sensing, UAV remote sensing and ground sensor data sources, and record them in real time through a data recorder; The auxiliary data unit is used to collect basic geographic information, mining plans and historical monitoring records of the mining area in real time through a data receiver and a data recorder.

3. The method according to claim 2, characterized in that: The data analysis module includes a data acquisition unit and an abnormality determination unit; The data acquisition unit is used to receive the basic geographic information, mining plan, historical monitoring records and current monitoring records of the mining area in real time through the data receiver; The abnormality determination unit is used to determine in real time whether the data monitored and recorded at the current moment needs to be corrected by using a calculation formula for the remote sensing data at the current moment. The monitoring record at the current moment includes the atmospheric factor correction value, radiation correction value and geometric correction value of the mine feature. First, the vegetation index calculation formula considering the atmospheric influence is used by the enhanced mine feature vegetation index 2, which is: enhanced mine feature vegetation index 2 = 2.5*(NIR–RED) / (NIR+2.4*RED+1)*(1-ATB); Among them, NIR is the near-infrared band reflectance, RED is the red band reflectance, and ATB is the atmospheric correction factor; The calculation formula for determining the abnormal data of the monitoring record at the current moment is as follows: Among them, P n represents the judgment data value, U1 represents the data value before correction of the current monitoring record, U2 represents the data value after correction of the current monitoring record, x represents the standard remote sensing data energy consumption of the data value, and X represents the standard remote sensing data energy consumption of the corrected data value. If the calculated E n = 0, the data does not need to be corrected. If E n ≠0, it is determined that the data needs to be corrected.

4. The method according to claim 3, characterized in that: The data tracking module includes a data tracking unit and an abnormal warning unit; The data tracking unit is used to track the monitoring record data value at the current moment in real time through a data tracker; The abnormal warning unit is used to send a voice alarm to the reporting system when it is determined that the data needs to be corrected, and to record the abnormal data that needs to be corrected in real time through color identification.

5. The method according to claim 1, characterized in that: The data fusion end includes a data processing module, a data fusion module and a model building module; The data processing module includes a data correction unit, which is used to correct the heterogeneous data that is determined to need correction. The correction formula is as follows: Among them, U, V and x, y are the image coordinates of the data value before and after the transformation, P ij and Q ij are the polynomial correction coefficients, n represents the polynomial degree, and the calculated x and y represent the geometric correction data values.

6. The method according to claim 5, characterized in that: The data fusion module includes an image generation unit and an image extraction unit; The image generating unit is used to automatically generate remote sensing images through image fusion technology; The image extraction unit is used to determine what the mine feature area on the remote sensing image is based on the mine feature shape features on the remote sensing image. The determination steps are as follows: According to the first point of the mine feature shape on the remote sensing image, the equidistant feature points of the same shape of the mine feature and the currently tracked feature points of the mine feature shape, the average distance of the equidistant feature points of the mine feature shape is calculated. The currently tracked feature points are the feature point positions of the remote sensing image in different time periods. The average distance of the feature points is the distance between the mine feature in different time periods and the standard mine feature area of ​​the remote sensing image. Then, according to the different point values ​​of the feature points at different times, the approximate outline of the mine feature in the current remote sensing image is calculated, and the average distance Z of the matching feature points is calculated. t The calculation formula is as follows: Among them, f x,y-1 represents the characteristic points of the shape of the mine features in the specified area of ​​the remote sensing image at time t-1, f x,y represents the feature point monitored and recorded at time t, M represents the distance between the feature points tracked at time t, x represents the value of a point in the X-axis direction, and y represents the value of the moving position in the Y-axis direction; By moving the pixel point on the known path at time t-1 by Z t , the approximate outline of the objects in the mining features of the remote sensing image at time t can be obtained.

7. The method according to claim 6, characterized in that: The model building module includes a model building unit, a manual interpretation unit, an online monitoring unit and a trend warning unit; The model building unit is used to construct a convolutional neural network model, which is used to set mine feature identification parameters based on multi-source remote sensing data fusion technology, and the mine feature identification parameters include standard values ​​of environmental parameters, disaster warning parameter levels, mineral distribution range parameters, vegetation destruction parameters and soil and water loss parameters; The manual interpretation unit is used to introduce manual visual interpretation as an auxiliary means to verify and correct the results of machine interpretation; The online monitoring unit is used to monitor the basic parameters of the current mining area in real time through data monitoring equipment. The basic parameters include environmental parameters, disaster level, mineral distribution range, vegetation destruction rate, soil erosion rate and historical parameter difference; The trend warning unit is used to create, manage, analyze and draw each remote sensing image area through GIS, and predict the development trend of mining objects in real time through the mining object monitoring parameters in the remote sensing image area. The prediction method is as follows: Step I: Set the planning standard parameters for the mine feature planning project in the corresponding remote sensing image area, use earthquake monitoring equipment, landslide monitoring equipment, debris flow monitoring equipment, ground subsidence monitoring equipment and volcanic eruption monitoring equipment, temperature sensors, humidity sensors and light sensors to detect the execution parameters of the corresponding mine feature planning area in real time, and analyze the environmental parameters, disaster level, mineral distribution range, vegetation destruction rate, soil erosion rate and historical parameter difference of the corresponding mine feature planning area in real time; Step II: Calculate the difference between the execution parameters of the corresponding area and the planned standard parameters. If the difference = ±0.002, it is judged that the planned parameters tend to the safe value. If the difference > 0.002 or < -0.002, it is judged that the planned parameters exceed the safe value, and a planned trend warning signal is issued. Seven days is a cycle. The development trend of the mining features in the corresponding area is predicted based on the planned parameters of three cycles. If the average values ​​of the planned parameters of the three cycles do not exceed the planned standard parameters, the mining feature development plan of the corresponding area is judged to be safe and stable. If the average values ​​of the planned parameters of the three cycles exceed the planned standard parameters, the mining feature development plan of the corresponding area is judged to be abnormal.

8. The method according to claim 1, characterized in that: The comprehensive recognition end includes an image segmentation module, a feature extraction module, a feature classification module and a comprehensive recognition module; The image segmentation module includes a land feature type unit, which is used to divide the mine land feature area into different land feature types by calculating the approximate contour. The land feature types include mine pits, waste rock piles and tailings ponds.

9. The method according to claim 8, characterized in that: The feature extraction module is used to extract the spectrum, texture and shape features in the mining feature area through the calculation formula of the image extraction unit; The feature classification module is used to classify the extracted mine features into different categories, including mine pits, waste rock piles, rivers, buildings and tailings ponds.

10. The method according to claim 9, characterized in that: The comprehensive analysis module includes a multi-source data combination unit and a data verification unit; The multi-source data combination unit is used to combine remote sensing data with GIS, use the spatial analysis function of GIS to perform spatial distribution and statistical analysis on mining objects, and combine remote sensing data with ground survey data to perform difference calculation between the data analysis value at the current moment and the trend warning value in real time. If the difference is 0, the data analysis is judged to be valid, and if the difference is ≠ 0, the data analysis is judged to be invalid, and data correction is performed twice; The data verification unit is used to perform data verification in real time on the data results of the multi-source data structure unit through a data anomaly calculation formula to correct the data combination and data recognition results.

Images