A method and system for detecting dissolution of Danxia landform

Abstract

The present invention relates to the field of digital geological detection technology, and specifically, to a Danxia landform dissolution detection method and system, which mainly includes setting an initial collection interval, obtaining weather data for each group of basic data, grouping the basic data according to the weather data of the basic data, and obtaining a basic data set of several groups of different weather data; establishing a neural network prediction model, obtaining several neural network prediction model output values, and outputting the future dissolution degree of the Danxia landform according to the several neural network prediction model output values ​​through a comprehensive evaluation model. Through the above method, the influence of weather is taken as a consideration factor, and different weather data are grouped, so that data of different weather conditions are analyzed in the same group, and the detection error caused by weather factors is reduced as much as possible. At the same time, a neural network model is used to analyze the basic data, which further improves the accuracy of the prediction results based on the basic data.

Description

Technical Field

[0001] The present invention relates to the technical field of digital geological detection, and in particular to a Danxia landform dissolution detection method and system. Background Art

[0002] Danxia landform is a special red rock landform, mainly formed by sedimentary rocks such as sandstone and quartzite under geological action. The most notable feature of Danxia landform is its color, which is usually bright red, orange or yellow, due to the presence of iron oxides. The rock layers in Danxia landform show a distinct layered structure, usually horizontal or inclined sedimentary layers, with clear stratification, and sediments formed in different geological periods can be clearly seen. The formation of Danxia landform often undergoes long-term geological changes, and its geological history is rich in records, which is an important window for studying geological evolution.

[0003] At present, the detection of Danxia landforms includes field observation and measurement to record the morphological characteristics, cracks, erosion degree, etc. of Danxia landforms. If possible, drones can be used for aerial photography to obtain more comprehensive landform information. This includes a detailed investigation of the rock composition, structure, weathering layer, etc. of Danxia landforms, and analysis of their sensitivity to dissolution. Satellite remote sensing and aerial remote sensing technology are used to monitor a large range of Danxia landforms and analyze their changes, but the current existing technologies do not take into account the impact of weather on detection data. Weather factors (such as rainfall, drought, weathering, etc.) may affect the accuracy and real-time nature of detection, especially under extreme climatic conditions. The terrain and activities may change rapidly, so the data finally obtained may not accurately reflect the real situation. Summary of the invention

[0004] The purpose of the present invention is to provide a Danxia landform dissolution detection method and system to solve the above problems in the prior art.

[0005] The present invention is achieved through the following technical solutions:

[0006] In a first aspect, the present invention provides a method for detecting dissolution of Danxia landforms, comprising:

[0007] Acquire the first basic data of the target Danxia landform, and set an initial collection interval, and when the duration of the initial collection interval is reached, acquire the second basic data of the target Danxia landform;

[0008] Establishing a collection cycle correction model, correcting the initial collection interval by using the first basic data and the second basic data according to the collection cycle correction model, outputting the collection interval for the next collection, setting the number of collection times, and obtaining several groups of basic data;

[0009] Obtain weather data for each set of basic data, group the basic data according to the weather data of the basic data, and obtain basic data sets of several sets of different weather data;

[0010] Preprocess the basic data and extract features from the basic data, randomly select a number of basic data from each basic data set to form a feature subset, and divide the feature subset into a training set and a test set;

[0011] Establish a neural network prediction model, optimize the training calculation of the neural network prediction model through the training set, and optimize the trained neural network prediction model through the test set, update the weight, and output the optimal neural network prediction model after reaching the number of iterations;

[0012] The remaining feature subsets of each basic data set are evaluated by the neural network prediction model according to different groups to obtain several output values ​​of the neural network prediction model;

[0013] A comprehensive evaluation model is established, through which the future dissolution degree of Danxia landform is output according to the output values ​​of several neural network prediction models.

[0014] Preferably, obtaining basic data of the target Danxia landform includes:

[0015] Acquire geological data and remote sensing observation image data, and determine whether the remote sensing observation image data is blocked;

[0016] If there is no obstruction, no processing is performed;

[0017] If there is occlusion, the area of ​​the occluded region is obtained, and the proportion of the occluded region is obtained by the area of ​​the occluded region and the entire observed image data;

[0018] The proportion is used as a correction factor to correct the acquisition cycle correction model.

[0019] Preferably, the acquisition cycle correction model includes:

[0020]

[0021] In the formula, is the collection interval for the next collection, is the initial collection interval, is the soil thickness of the first basic data, is the soil thickness of the second basic data, The pH value of the first basic data is is the pH value of the second basic data, is the temperature of the first basic data, is the temperature of the second basic data, The humidity of the first basic data, The humidity of the second basic data, The area of ​​Danxia landform is the first basic data. The second basic data is the Danxia landform area. The number of people active in the Danxia landform area is the first basic data. The number of people active in the Danxia landform area is the second basic data. When the remote sensing observation image data is not blocked, , when the remote sensing observation image data is blocked, is the correction factor.

[0022] Preferably, grouping the basic data according to the weather data of the basic data comprises:

[0023] Obtain the meteorological data of the current basic data through the weather forecast system, and establish a weather classification model. Through the weather classification model and meteorological data, output the evaluation value of the current basic data;

[0024] Set the evaluation value threshold, and divide the current basic data into sunny day group and rainy day group according to the evaluation value and the evaluation value threshold.

[0025] The weather partitioning model includes:

[0026]

[0027] In the formula, is the evaluation value, is a variable coefficient. If the meteorological data indicates rain, then , if the meteorological data shows no rain, then , is the current soil moisture, is the soil moisture of the previous day, is the current rainfall probability.

[0028] Preferably, the evaluation threshold is:

[0029]

[0030] In the formula, is the evaluation value threshold, when Less than or equal to , the current basic data is divided into rainy day groups, when Greater than , the current basic data is divided into sunny day group.

[0031] Preferably, the preprocessing of the basic data includes:

[0032] Normalize the basic data:

[0033]

[0034] In the formula, is the normalized value, is the original value of the underlying data, is the maximum original value of the underlying data, The minimum raw value of the underlying data.

[0035] Preferably, the establishing of a neural network prediction model comprises:

[0036] Set the input layer, hidden layer, and output layer of the automobile export assessment model, and determine the weight coefficients between the input layer and the hidden layer, and between the hidden layer and the output layer:

[0037]

[0038] In the formula, Output value for the neural network prediction model, is the first weight matrix, is the second weight matrix, is the first bias vector, is the activation function, is the regularization coefficient, is the number of training sets, For the The input features of the training set samples, is the second bias vector, For the training set samples to The output of the hidden layer is is the learning rate, For the hidden layers to The output of the output layer, For the Input layer to The output of the hidden layer is The input layer Features, The hidden layer Features, The output layer Features, For the Input layer to The output of the hidden layer is is the time period to be predicted, , , , , , , All are natural numbers.

[0039] Preferably, the comprehensive evaluation model includes:

[0040]

[0041] In the formula, is the evaluation value, is the number of output values ​​of the neural network prediction model, For the A neural network predicts the model output value.

[0042] Preferably, the future dissolution degree of the output Danxia landform includes:

[0043] Obtaining the evaluation value of the comprehensive evaluation model;

[0044] If the evaluation value is greater than 0.5, the output is The result of the gradual accelerated dissolution of Danxia landforms in the future within the time period;

[0045] If the evaluation value is less than 0.5, the output is The result of the gradual slowdown of dissolution of Danxia landforms in the future within the time period;

[0046] If the evaluation value is equal to 0.5, then output The result is that the Danxia landform will remain unchanged in the future within the time period.

[0047] In a second aspect, the present invention further provides a Danxia landform dissolution detection system, comprising:

[0048] The data acquisition module is configured to acquire the first basic data of the target Danxia landform and set an initial acquisition interval. When the duration of the initial acquisition interval is reached, the second basic data of the target Danxia landform is acquired; an acquisition cycle correction model is established, and the initial acquisition interval is corrected by the first basic data and the second basic data according to the acquisition cycle correction model, and the acquisition interval of the next acquisition is output, and the number of acquisitions is set to obtain several groups of basic data; the weather data of each group of basic data is acquired, and the basic data is grouped according to the weather data of the basic data to obtain several groups of basic data sets with different weather data;

[0049] The neural network prediction module is configured to pre-process the basic data and extract features from the basic data, randomly select a number of basic data from each basic data set to form a feature subset, and divide the feature subset into a training set and a test set; establish a neural network prediction model, optimize the training calculation of the neural network prediction model through the training set, and optimize the trained neural network prediction model through the test set, update the weight, and output the optimal neural network prediction model after reaching the number of iterations; evaluate the remaining feature subsets of each basic data set through the neural network prediction model according to different groups, and obtain a number of neural network prediction model output values;

[0050] An output module is configured to establish a comprehensive evaluation model, and output the future dissolution degree of the Danxia landform according to the output values ​​of several neural network prediction models through the comprehensive evaluation model;

[0051] A main control module, which is connected to the data acquisition module, the neural network prediction module and the output module, is used to execute the above-mentioned Danxia landform dissolution detection method.

[0052] The technical solution of the embodiment of the present invention has at least the following advantages and beneficial effects:

[0053] The method provided by the present invention mainly includes setting the initial collection interval, correcting the initial collection interval through the first basic data and the second basic data according to the collection cycle correction model, obtaining the weather data of each group of basic data, grouping the basic data according to the weather data of the basic data, and obtaining a basic data set of several groups of different weather data; establishing a neural network prediction model, obtaining several neural network prediction model output values, and outputting the future dissolution degree of Danxia landform according to several neural network prediction model output values ​​through a comprehensive evaluation model. Through the above method, the influence of weather is taken as a consideration, and different weather data are grouped, and the basic data of the target Danxia landform belonging to the same group are centrally analyzed and calculated, and then a comprehensive calculation is performed, so that the data of different weather conditions are analyzed in the same group, with high comparability, and the detection error caused by weather factors is reduced as much as possible. At the same time, a neural network model is used to analyze the basic data, which further improves the accuracy of the prediction results based on the basic data, provides relatively accurate information to the staff, and responds. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0055] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0057] The terms "first, second" and the like in the specification and claims of this application and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. The naming or numbering of steps in this application does not mean that the steps in the method flow must be executed in the time / logical sequence indicated by the naming or numbering. The named or numbered process steps can be changed in the execution order according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved.

[0058] The modules or submodules described independently may be physically separated or not: they may be implemented by software or hardware, and some modules or submodules may be implemented by software, and the processor may call the software to implement the functions of these modules or submodules, and other modules or submodules may be implemented by hardware, such as by hardware circuits. In addition, some or all of the modules may be selected according to actual needs to achieve the purpose of the present application.

[0059] Please refer to Figure 1 The present invention provides a method for detecting Danxia landform dissolution, comprising:

[0060] S101: Acquire first basic data of the target Danxia landform, and set an initial acquisition interval, and when the duration of the initial acquisition interval is reached, acquire second basic data of the target Danxia landform;

[0061] In this embodiment, the basic data includes geological data and remote sensing observation image data. The geological data includes groundwater level, permeability, drainage pattern, etc., which affect the flow of water and the dissolution process of Danxia landform. Soil composition, properties (such as pH, organic matter content, water retention capacity, etc.) and their interaction with rocks. The elevation, slope and orientation data of the terrain can reflect the intensity of weathering and erosion and their impact on the landform.

[0062] The remote sensing observation image data mainly include remote sensing observation images, through which the overall appearance changes of the Danxia landform can be more intuitively observed.

[0063] S102: Establishing a collection cycle correction model, correcting the initial collection interval by using the first basic data and the second basic data according to the collection cycle correction model, outputting the collection interval for the next collection, setting the number of collection times, and obtaining several groups of basic data;

[0064] In the present invention, data is not collected at a fixed collection interval, which may cause data lag or waste of other resources. The collection interval should be set within a reasonable range. Therefore, in this embodiment, a variable collection interval is used to determine the next collection interval based on two specific adjacent basic data. In this way, the next collection interval is as close to a reasonable range as possible, reducing collection lag or avoiding waste caused by too frequent collection.

[0065] S103: Acquire weather data of each group of basic data, group the basic data according to the weather data of the basic data, and obtain a number of basic data sets of different weather data;

[0066] In this embodiment, the meteorological data of the current basic data is obtained through the weather forecast system, and a weather classification model is established. The evaluation value of the current basic data is output through the weather classification model and the meteorological data; the evaluation value threshold is set, and the current basic data is divided into a sunny day group and a rainy day group through the evaluation value and the evaluation value threshold. That is, the basic data of this embodiment is divided into two groups of basic data sets, and the basic data of the same group are separately analyzed and calculated.

[0067] S104: preprocessing the basic data, extracting features from the basic data, randomly selecting a number of basic data from each basic data set to form a feature subset, and dividing the feature subset into a training set and a test set;

[0068] In this implementation, conventional technical means in the prior art may be used for feature extraction, and feature extraction may be performed based on the importance of data.

[0069] S105: Establishing a neural network prediction model, optimizing the training calculation of the neural network prediction model through the training set, optimizing the trained neural network prediction model through the test set, updating the weight, and outputting the optimal neural network prediction model after reaching the number of iterations;

[0070] S106: Evaluate the remaining feature subsets of each basic data set by using a neural network prediction model according to different groups to obtain a number of neural network prediction model output values;

[0071] S107: Establish a comprehensive evaluation model, and output the future dissolution degree of the Danxia landform according to the output values ​​of several neural network prediction models through the comprehensive evaluation model.

[0072] The method provided by the present invention mainly includes setting the initial collection interval, correcting the initial collection interval through the first basic data and the second basic data according to the collection cycle correction model, obtaining the weather data of each group of basic data, grouping the basic data according to the weather data of the basic data, and obtaining a basic data set of several groups of different weather data; establishing a neural network prediction model, obtaining several neural network prediction model output values, and outputting the future dissolution degree of Danxia landform according to several neural network prediction model output values ​​through a comprehensive evaluation model. Through the above method, the influence of weather is taken as a consideration, and different weather data are grouped, and the basic data of the target Danxia landform belonging to the same group are centrally analyzed and calculated, and then a comprehensive calculation is performed, so that the data of different weather conditions are analyzed in the same group, with high comparability, and the detection error caused by weather factors is reduced as much as possible. At the same time, a neural network model is used to analyze the basic data, and the prediction structure based on the detection results is further improved, providing relatively accurate information to the staff, and responding.

[0073] In an exemplary embodiment of the present invention, obtaining basic data of target Danxia landform includes:

[0074] Obtain geological data and remote sensing observation image data, and determine whether the remote sensing observation image data is blocked; if not, do not process it; if there is occlusion, obtain the area of ​​the blocked area, and obtain the proportion of the blocked area by the area of ​​the blocked area and the entire observation image data; use the proportion as a correction factor to correct the acquisition period correction model.

[0075] In this embodiment, it is taken into consideration that the remote sensing observation image data may not fully reflect the actual situation, and the Danxia landform may also be obstructed by other obstructions. This part of the change cannot be fed back through the graphic data, so this part is converted into a correction factor to provide correction for the subsequent model.

[0076] Specifically, the collection cycle correction model includes:

[0077]

[0078] In the formula, is the collection interval for the next collection, is the initial collection interval, The soil thickness is the first basic data, is the soil thickness of the second basic data, The pH value of the first basic data is is the pH value of the second basic data, is the temperature of the first basic data, is the temperature of the second basic data, The humidity of the first basic data, The humidity of the second basic data, The area of ​​Danxia landform is the first basic data. The second basic data is the Danxia landform area. The number of people active in the Danxia landform area is the first basic data. The number of people active in the Danxia landform area is the second basic data. When the remote sensing observation image data is not blocked, , when the remote sensing observation image data is blocked, is the correction factor.

[0079] In an exemplary embodiment of the present invention, the weather classification model includes:

[0080]

[0081] In the formula, is the evaluation value, is a variable coefficient. If the meteorological data indicates rain, then , if the meteorological data shows no rain, then , is the current soil moisture, is the soil moisture of the previous day, is the current rainfall probability.

[0082] Specifically, the evaluation threshold is:

[0083]

[0084] In the formula, is the evaluation value threshold, when Less than or equal to , the current basic data is divided into rainy day groups, when Greater than , the current basic data is divided into sunny day group.

[0085] For example, the current soil moisture is 80%, the soil moisture the day before was 20%, the meteorological data shows no rain, and the current probability of rainfall is 60%, so we can get: , Therefore, it can be judged as the rainy day group, because the Danxia landform has a great connection with soil moisture. Even if there is no obvious rain, due to the huge change in soil moisture and the probability of rainfall, the basic data collected without obvious rain can be classified as the rainy day group to reduce subsequent analysis errors.

[0086] In an exemplary embodiment of the present invention, preprocessing the basic data includes:

[0087] Normalize the basic data:

[0088]

[0089] In the formula, is the normalized value, is the original value of the underlying data, is the maximum original value of the underlying data, The minimum raw value of the underlying data.

[0090] In an exemplary embodiment of the present invention, establishing a neural network prediction model includes:

[0091] Set the input layer, hidden layer, and output layer of the automobile export assessment model, and determine the weight coefficients between the input layer and the hidden layer, and between the hidden layer and the output layer:

[0092]

[0093] In the formula, Output value for the neural network prediction model, is the first weight matrix, is the second weight matrix, is the first bias vector, is the activation function, is the regularization coefficient, is the number of training sets, For the The input features of the training set samples, is the second bias vector, For the training set samples to The output of the hidden layer is is the learning rate, For the hidden layers to The output of the output layer, For the Input layer to The output of the hidden layer is The input layer Features, The hidden layer Features, The output layer Features, For the Input layer to The output of the hidden layer is is the time period to be predicted, , , , , , , All are natural numbers.

[0094] The optimization and training process may include calculating the neural network prediction model through the training set;

[0095] Increase the values ​​of weights and bias terms, and calculate a squared error loss value for each weight and bias update. If the squared error loss value decreases as the weight increases, then the weight value is increased. The amount of increase in the weight value should be proportional to the amount of loss value reduced by the weight change;

[0096] Loop through all weight values ​​and then make the minimum change to them. After the weights are modified, calculate the updated forward propagation loss;

[0097] Update the weights by changing the loss, determine the learning rate parameters, and return the updated weights and bias values ​​until the number of iterations is reached.

[0098] In an exemplary embodiment of the present invention, the comprehensive evaluation model includes:

[0099]

[0100] In the formula, is the evaluation value, is the number of output values ​​of the neural network prediction model, For the A neural network predicts the model output value.

[0101] In an exemplary embodiment of the present invention, the future dissolution degree of the output Danxia landform includes:

[0102] Obtaining the evaluation value of the comprehensive evaluation model;

[0103] If the evaluation value is greater than 0.5, the output is The result of the gradual accelerated dissolution of Danxia landforms in the future within the time period;

[0104] If the evaluation value is less than 0.5, the output is The result of the gradual slowdown of dissolution of Danxia landforms in the future within the time period;

[0105] If the evaluation value is equal to 0.5, then output The result is that the Danxia landform will remain unchanged in the future within the time period.

[0106] The present invention also provides a Danxia landform dissolution detection system, comprising:

[0107] The data acquisition module is configured to acquire the first basic data of the target Danxia landform and set an initial acquisition interval. When the duration of the initial acquisition interval is reached, the second basic data of the target Danxia landform is acquired; an acquisition cycle correction model is established, and the initial acquisition interval is corrected by the first basic data and the second basic data according to the acquisition cycle correction model, and the acquisition interval of the next acquisition is output, and the number of acquisitions is set to obtain several groups of basic data; the weather data of each group of basic data is acquired, and the basic data is grouped according to the weather data of the basic data to obtain several groups of basic data sets with different weather data;

[0108] The neural network prediction module is configured to pre-process the basic data and extract features from the basic data, randomly select a number of basic data from each basic data set to form a feature subset, and divide the feature subset into a training set and a test set; establish a neural network prediction model, optimize the training calculation of the neural network prediction model through the training set, and optimize the trained neural network prediction model through the test set, update the weight, and output the optimal neural network prediction model after reaching the number of iterations; evaluate the remaining feature subsets of each basic data set through the neural network prediction model according to different groups, and obtain a number of neural network prediction model output values;

[0109] An output module is configured to establish a comprehensive evaluation model, and output the future dissolution degree of the Danxia landform according to the output values ​​of several neural network prediction models through the comprehensive evaluation model;

[0110] A main control module, which is connected to the data acquisition module, the neural network prediction module and the output module, is used to execute the above-mentioned Danxia landform dissolution detection method.

[0111] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0112] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0113] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for detecting dissolution of Danxia landform, characterized in that: include: Acquire the first basic data of the target Danxia landform, and set an initial acquisition interval. When the duration of the initial acquisition interval is reached, acquire the second basic data of the target Danxia landform, the basic data including geological data and remote sensing observation image data; Establishing a collection cycle correction model, correcting the initial collection interval by using the first basic data and the second basic data according to the collection cycle correction model, outputting the collection interval for the next collection, setting the number of collection times, and obtaining several groups of basic data; The acquisition cycle correction model includes: In the formula, is the collection interval for the next collection, is the initial collection interval, The soil thickness is the first basic data, is the soil thickness of the second basic data, The pH value of the first basic data is is the pH value of the second basic data, is the temperature of the first basic data, is the temperature of the second basic data, The humidity of the first basic data, The humidity of the second basic data, The area of ​​Danxia landform is the first basic data. The second basic data is the Danxia landform area. The number of people active in the Danxia landform area is the first basic data. The number of people active in the Danxia landform area is the second basic data. When the remote sensing observation image data is not blocked, , when the remote sensing observation image data is blocked, is a correction factor; obtaining weather data for each set of basic data, grouping the basic data according to the weather data of the basic data, and obtaining a set of basic data for several sets of different weather data; The grouping of the basic data according to the weather data of the basic data comprises: Obtain the meteorological data of the current basic data through the weather forecast system, and establish a weather classification model. Through the weather classification model and meteorological data, output the evaluation value of the current basic data; Set the evaluation value threshold, and divide the current basic data into sunny day group and rainy day group according to the evaluation value and the evaluation value threshold; The weather partitioning model includes: In the formula, is the evaluation value, is a variable coefficient. If the meteorological data indicates rain, then , if the meteorological data shows no rain, then , is the current soil moisture, is the soil moisture of the previous day, is the current rainfall probability; Preprocess the basic data and extract features from the basic data, randomly select a number of basic data from each basic data set to form a feature subset, and divide the feature subset into a training set and a test set; Establish a neural network prediction model, optimize the training calculation of the neural network prediction model through the training set, and optimize the trained neural network prediction model through the test set, update the weight, and output the optimal neural network prediction model after reaching the number of iterations; The remaining feature subsets of each basic data set are evaluated by the neural network prediction model according to different groups to obtain several output values ​​of the neural network prediction model; A comprehensive evaluation model is established, through which the future dissolution degree of Danxia landform is output according to the output values ​​of several neural network prediction models.

2. A Danxia landform dissolution detection method according to claim 1, characterized in that: The basic data for obtaining the target Danxia landform include: Acquire geological data and remote sensing observation image data, and determine whether the remote sensing observation image data is blocked; If there is no obstruction, no processing is performed; If there is occlusion, the area of ​​the occluded region is obtained, and the proportion of the occluded region is obtained by the area of ​​the occluded region and the entire observed image data; The proportion is used as a correction factor to correct the acquisition cycle correction model.

3. A Danxia landform dissolution detection method according to claim 2, characterized in that: The evaluation value threshold is: In the formula, is the evaluation value threshold, when Less than or equal to , the current basic data is divided into rainy day groups, when Greater than , the current basic data is divided into sunny day group.

4. A Danxia landform dissolution detection method according to claim 3, characterized in that: The preprocessing of basic data includes: Normalize the basic data: In the formula, is the normalized value, is the original value of the underlying data, is the maximum original value of the underlying data, The minimum raw value of the underlying data.

5. A Danxia landform dissolution detection method according to claim 4, characterized in that: The establishment of a neural network prediction model comprises: Set the input layer, hidden layer, and output layer of the automobile export assessment model, and determine the weight coefficients between the input layer and the hidden layer, and between the hidden layer and the output layer: In the formula, Output value for the neural network prediction model, is the first weight matrix, is the second weight matrix, is the first bias vector, is the activation function, is the regularization coefficient, is the number of training sets, For the The input features of the training set samples, is the second bias vector, For the training set samples to The output of the hidden layer is is the learning rate, For the hidden layers to The output of the output layer, For the Input layer to The output of the hidden layer is The input layer Features, The hidden layer Features, The output layer Features, For the Input layer to The output of the hidden layer is is the time period to be predicted, , , , , , , All are natural numbers.

6. A Danxia landform dissolution detection method according to claim 5, characterized in that: The comprehensive assessment model includes: In the formula, is the evaluation value, is the number of output values ​​of the neural network prediction model, For the A neural network predicts the model output value.

7. A Danxia landform dissolution detection method according to claim 6, characterized in that: The future dissolution degree of the output Danxia landform includes: Obtaining the evaluation value of the comprehensive evaluation model; If the evaluation value is greater than 0.5, the output is The result of the gradual accelerated dissolution of Danxia landforms in the future within the time period; If the evaluation value is less than 0.5, the output is The result of the gradual slowdown of dissolution of Danxia landforms in the future within the time period; If the evaluation value is equal to 0.5, then output The result is that the Danxia landform will remain unchanged in the future within the time period.

Images