CN114596498B - Assignment method and system for geochemical sampling blind area and storage medium - Google Patents
Assignment method and system for geochemical sampling blind area and storage medium Download PDFInfo
- Publication number
- CN114596498B CN114596498B CN202210503858.4A CN202210503858A CN114596498B CN 114596498 B CN114596498 B CN 114596498B CN 202210503858 A CN202210503858 A CN 202210503858A CN 114596498 B CN114596498 B CN 114596498B
- Authority
- CN
- China
- Prior art keywords
- sampling
- area
- blind area
- target
- virtual
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a method, a system and a storage medium for assigning a geochemical sampling blind area, which relate to the technical field of geochemical evaluation, and the method comprises the following steps: delineating a sampling blind area in the area to be assigned based on a sample acquisition result of the area to be assigned; predicting the geochemical data of the sampling blind area by combining the sample acquisition result and the geographic factors of the area to be assigned, wherein the geographic factors comprise geological factors and geospatial information; assigning the geochemical data to the sampling dead zone. The method and the device have the effect of relatively accurate geochemical data predicted for the sampling blind area.
Description
Technical Field
The application relates to the technical field of geochemistry evaluation, in particular to a method and a system for assigning a geochemistry sampling blind area and a storage medium.
Background
The land quality geochemistry evaluation unit is a minimum unit of an evaluation object and is a minimum space unit for dividing the soil nutrient geochemistry grade, the soil environment geochemistry grade and the land quality geochemistry comprehensive grade. Due to cost limitation and different precision requirements, in actual work, investigation and sampling of each pattern spot of the whole evaluation area cannot be achieved, so that the problem that the land quality geochemical investigation sampling result is not completely matched with the distribution of the pattern spots exists, and blank pattern spots which are not sampled are generally called sampling blind areas.
In the related technology, the element content in the soil is generally considered to be linearly changed in space, so that sampling is carried out in a random sampling mode, assignment is carried out on a pattern spot where a sampling point is located according to a sampling result, and finally linear simulation assignment is carried out on a sampling blind area where sampling is not carried out according to the linear change trend of the element content.
With respect to the related art among the above, the inventors consider that the following drawbacks exist: the linear simulation assignment mode only analyzes the linear relation of the element content in the soil, if the element distribution variability of the evaluation area is large, the element content in the soil of the evaluation area does not have obvious linear relation, and at the moment, the assignment result of assigning the blank pattern spots by adopting the linear simulation assignment mode has large error.
Disclosure of Invention
In order to overcome the defect that an assignment result obtained by assigning a blank pattern spot in a linear simulation assignment mode has a large error when element distribution variability in an evaluation area is large, the application provides an assignment method, an assignment system and a storage medium for a geochemical sampling dead zone.
In a first aspect, the present application provides a method for assigning a geochemical sampling dead zone, the method comprising the steps of:
delineating a sampling blind area in the area to be assigned based on a sample acquisition result of the area to be assigned;
predicting the geochemical data of the sampling blind area by combining the sample acquisition result and the geographic factors of the area to be assigned, wherein the geographic factors comprise geological factors and geospatial information;
assigning the geochemical data to the sampling blind area.
By adopting the technical scheme, the sampling blind area which is not sampled is defined according to the sample acquisition result which is randomly sampled in the area to be assigned, and the geological factors and the geospatial information of the whole area to be assigned are analyzed in the process of predicting the geochemical data of the sampling blind area according to the sample acquisition result, so that the predicted geochemical data is assigned to the sampling blind area which is not sampled.
Optionally, the predicting the geochemical data of the sampling dead zone by combining the sample acquisition result and the geographic factor of the area to be assigned comprises the following steps:
calculating the similarity between the sampling blind area and a sample acquisition point based on the geographic factors of the area to be assigned;
judging whether the similarity exceeds a preset similarity threshold value or not;
if the similarity exceeds the similarity threshold, marking the corresponding sample acquisition point as a target sample acquisition point;
respectively configuring the assignment weights of all target sample acquisition points according to the similarity;
and predicting the geochemical data of the sampling dead zone by combining the sample acquisition results of all target sample acquisition points and the assigned weights.
By adopting the technical scheme, because the distribution of the plurality of sample acquisition points is different, the difference between the geographic factors of the positions of the sample acquisition points and the geographic factors of the sampling blind areas is different, the similarity between each sample acquisition point and the sampling blind area is calculated firstly, then the sample acquisition points with high similarity are screened out through a preset similarity threshold value to be used as target sample acquisition points, different assignment weights are configured for all the target sample acquisition points according to the similarity, and finally the assignment weights of the target sample acquisition points and the sample acquisition results are combined for calculation, so that the geochemical data of the sampling blind areas are predicted.
Optionally, the calculating the similarity between the sampling blind area and the sample collection point based on the geographic factor of the to-be-assigned area includes the following steps:
calculating a geographic factor difference value between the sampling blind area and the sample collection point based on geographic factors of the sampling blind area and the sample collection point;
configuring an initial weight value for the geographic factor difference value;
optimizing the initial weight value through a preset group intelligent algorithm to obtain a final weight value;
and calculating the similarity between the sampling blind area and the sample collection point by combining the final weight value and the geographic factor difference value.
By adopting the technical scheme, when the similarity between the sampling blind area and the sample collection point is calculated, a plurality of geographic factors of the sampling blind area and the sample collection point need to be considered, the geographic factor difference value of each geographic factor can be calculated firstly, the initial weight value is configured for the geographic factor difference value, the initial weight value is optimized to be the final weight value through the group intelligent algorithm, the calculation of the similarity is more accurate, and the similarity between the sampling blind area and the sample collection point is calculated finally according to the geographic factor difference value and the corresponding final weight value.
Optionally, the pattern spot in the to-be-assigned region where the sample acquisition result is located is a sampling region, and predicting the geochemical data of the sampling blind region by combining the sample acquisition result and the geographic factor of the to-be-assigned region includes the following steps:
generating adjacency relations between all sampling areas and all sampling blind areas based on the geospatial information of the area to be assigned;
predicting geochemical data of a target sampling blind area adjacent to the sampling area based on the adjacency relation and according to the sample acquisition result;
judging whether the target sampling blind area has other adjacent sampling blind areas or not based on the adjacency relation;
if the target sampling blind area has other adjacent sampling blind areas, predicting the geochemical data of the other sampling blind areas based on the geochemical data of the target sampling blind area;
marking the other sampling blind areas as the target sampling blind areas;
and repeating the judging step until all the sampling dead zones in the to-be-assigned zone are marked as the target sampling dead zone.
By adopting the technical scheme, the adjacency relation between the sampling area and the sampling blind area is generated according to the geospatial information of each area in the area to be assigned, the two areas are adjacent to each other and represent similar geospatial information between the two areas, and the sampling area is the area for sampling and obtaining the sampling result, so that the geochemical data of the adjacent target sampling blind area can be predicted according to the sampling result of the sampling area, and after the geochemical data of the target sampling blind area is obtained through prediction, the geochemical data of other sampling blind areas adjacent to the target sampling blind area can be predicted continuously according to the geochemical data of the target sampling blind area, so that the data prediction of all sampling blind areas in the whole area to be assigned is completed through a data transmission prediction mode.
Optionally, the predicting the geochemical data of the target sampling blind area adjacent to the sampling area based on the adjacency relation and according to the sample acquisition result comprises the following steps:
calculating feature similarity between all adjacent regions based on the geological factors;
judging whether a target sampling blind area adjacent to the sampling area is adjacent to other sampling areas;
if the target sampling blind area is not adjacent to other sampling areas, predicting the geochemical data of the target sampling blind area according to the sample acquisition result of the sampling area;
if the target sampling blind area is adjacent to other sampling areas, configuring different predicted value weights for all adjacent sampling areas respectively based on the feature similarity of the target sampling blind area and all adjacent sampling areas;
and predicting the geochemical data of the target sampling blind area by combining the sample acquisition results and the predicted value weights of all adjacent sampling areas.
By adopting the technical scheme, after the adjacency relation is generated according to the geographic space information, the characteristic similarity of the geological characteristics between the two adjacent regions can be calculated according to the geological factors of the two adjacent regions, the number of the sampling regions adjacent to the target sampling blind region needs to be judged before the data prediction of the target sampling blind region adjacent to the sampling region, and if the target sampling blind region is only adjacent to one sampling region, the geochemical data of the target sampling blind region is directly predicted according to the sample acquisition result of the sampling region; if the target sampling blind area is adjacent to the plurality of sampling areas, different predicted value weights need to be configured for the plurality of adjacent sampling areas according to the feature similarity between the target sampling blind area and the plurality of adjacent sampling areas, the higher the feature similarity is, the larger the configured predicted value weight is, and finally, the geochemical data of the target sampling blind area is predicted by combining the predicted value weights corresponding to the adjacent sampling areas and the sample acquisition results.
Optionally, the predicting the geochemical data of the sampling dead zone by combining the sample acquisition result and the geographic factor of the area to be assigned comprises the following steps:
fitting all the sample acquisition results through a preset deep learning model for training to obtain a plurality of virtual sampling points and virtual acquisition results corresponding to the virtual sampling points;
judging whether the virtual acquisition result is true or not based on the sample acquisition result;
if the virtual acquisition result is not true, feeding back a judgment result to the deep learning model;
if the virtual acquisition result is true, feeding a judgment result back to the deep learning model, and marking a target virtual sampling point corresponding to the virtual acquisition result in the area where the target virtual sampling point is located;
counting the sampling number of the target virtual sampling points in the sampling blind area;
and predicting the geochemical data of the sampling blind area by combining the virtual acquisition result of the target virtual sampling point and the sampling quantity.
By adopting the technical scheme, the collected actual sample collection result is used as the training basis of the deep learning model, a plurality of virtual collection results are obtained through training, the truth of the virtual collection result is judged according to the sample collection result, the judgment result is fed back to the deep learning model so as to optimize the deep learning model, and the virtual collection result generated subsequently by the deep learning model can be closer to the sample collection result. If the judgment result is true, the target virtual sampling points corresponding to the virtual acquisition results can be used as sample sampling points to be marked to the corresponding positions in the area to be assigned, so that the geochemical data of the target sampling blind area can be predicted according to all the marked target virtual sampling points in the sampling blind area and the corresponding virtual acquisition results.
Optionally, the method further includes the following steps after the preset deep learning model is used to fit all the sample acquisition results for training to obtain a plurality of virtual sampling points and virtual acquisition results corresponding to the plurality of virtual sampling points:
reversely decoding all the virtual acquisition results to obtain a sample training set;
inputting the sample training set into the deep learning model.
By adopting the technical scheme, the generated virtual acquisition result is reversely decoded into the sample training set, and the sample training set is input into the deep learning model, so that the training speed of the deep learning model can be accelerated, and the deep learning model can more quickly generate the virtual acquisition result close to the actual sample acquisition result.
Optionally, the predicting the geochemical data of the sampling dead zone by combining the virtual acquisition result of the target virtual sampling point and the sampling number comprises the following steps:
judging whether the sampling number exceeds a preset number threshold value;
if the sampling quantity exceeds the quantity threshold value, predicting the geochemical data of the sampling blind area by combining the virtual acquisition results of all the target virtual sampling points;
if the sampling quantity does not exceed the quantity threshold, processing the virtual acquisition results of all target virtual sampling points by an interpolation method to obtain a prediction curve;
and predicting the geochemical data of the sampling blind area according to the prediction curve.
By adopting the technical scheme, if the number of the virtual sampling points in the sampling blind area is small, the prediction result is possibly inaccurate, so that the sampling number of the virtual sampling points in the sampling blind area is judged through a preset number threshold, and when the sampling number is less than the number threshold, the virtual acquisition results of all the virtual sampling points in the sampling blind area need to be completed through an interpolation method so as to generate a more accurate prediction curve to predict the geochemical data of the sampling blind area.
In a second aspect, the present application further provides a assignment system for geochemical sampling dead zones, which includes a memory, a processor, and a program stored in the memory and executable on the processor, where the program is capable of being loaded and executed by the processor to implement the assignment method for geochemical sampling dead zones as described in the first aspect.
By adopting the technical scheme, through program calling, a sampling blind area which is not sampled is defined according to a sample acquisition result which is randomly sampled in the area to be assigned, and analysis on geological factors and geospatial information of the whole area to be assigned is added in the process of predicting the geochemical data of the sampling blind area according to the sample acquisition result, so that the predicted geochemical data is assigned to the sampling blind area which is not sampled.
In a third aspect, the present application further provides a computer-readable storage medium storing a computer program, which when executed by a processor causes the processor to implement the method for assigning a geochemical sampling dead zone according to the first aspect.
By adopting the technical scheme, through program calling, a sampling blind area which is not sampled is defined according to a sample acquisition result which is randomly sampled in the area to be assigned, and analysis on geological factors and geospatial information of the whole area to be assigned is added in the process of predicting the geochemical data of the sampling blind area according to the sample acquisition result, so that the predicted geochemical data is assigned to the sampling blind area which is not sampled.
To sum up, the application comprises the following beneficial technical effects:
the sampling dead zone which is not sampled is defined according to the sample acquisition result of random sampling in the area to be assigned, and the geological factors and the geographic space information of the whole area to be assigned are analyzed in the process of predicting the geochemical data of the sampling dead zone according to the sample acquisition result, so that the predicted geochemical data is assigned to the sampling dead zone which is not sampled.
Drawings
FIG. 1 is a schematic flow chart of a method for assigning geochemical dead sampling zones according to one embodiment of the present application.
FIG. 2 is a first schematic flow chart of geochemical data for a sample dead zone combined with sample acquisition and geodetic prediction, according to one embodiment of the present disclosure.
FIG. 3 is a schematic flow chart of calculating similarity between sample collection points of a sample dead zone box based on geographic factors according to an embodiment of the present application.
FIG. 4 is a schematic flow chart diagram of geochemical data for a sample dead zone combined with sample acquisition and geographic factor prediction according to one embodiment of the present application.
FIG. 5 is a schematic flow chart of the present application for predicting geochemical data for a target sample shadow adjacent a sample region based on adjacencies and sample acquisitions, according to one embodiment of the present application.
FIG. 6 is a third schematic flow chart of geochemical data for a sample dead zone combined with sample acquisition and geodetic prediction according to one embodiment of the present application.
FIG. 7 is a flowchart illustrating an optimized deep learning model according to an embodiment of the present application.
FIG. 8 is a schematic flow chart of geochemical data for predicting a sample dead zone in conjunction with a virtual acquisition of a target virtual sample point and a sample size, according to one embodiment of the present application.
Detailed Description
The present application is described in further detail below with reference to figures 1-8.
The embodiment of the application discloses an assignment method for geochemical sampling blind areas.
Referring to fig. 1, the assignment method of geochemical sampling dead zones comprises the following steps:
s101, defining sampling blind areas in the to-be-assigned areas based on sample collection results of the to-be-assigned areas.
The method comprises the steps of obtaining a sample acquisition result by randomly selecting points of a region to be assigned and sampling, wherein the pattern spots where random sampling points are located are sampling regions, and the pattern spots which do not contain the random sampling points are sampling blind regions.
And S102, predicting geochemical data of the sampling blind area by combining the sample acquisition result and the geographic factor of the area to be assigned.
The geographic factors comprise geological factors and geospatial information, and the geological factors comprise geological backgrounds, soil types and land utilization types.
And S103, assigning the geochemical data to a sampling blind area.
The implementation principle of the embodiment is as follows:
the sampling dead zone which is not sampled is defined according to the sample acquisition result of random sampling in the area to be assigned, and the geological factors and the geographic space information of the whole area to be assigned are analyzed in the process of predicting the geochemical data of the sampling dead zone according to the sample acquisition result, so that the predicted geochemical data is assigned to the sampling dead zone which is not sampled.
In step S102 of the embodiment shown in fig. 1, in one embodiment of this step, the similarity between the sampling blind area and the sample collection point is calculated according to the geographic factor of the to-be-assigned area, then the weight is configured according to the similarity, and finally the geochemical data of the sampling blind area is predicted. This is explained in detail with reference to the embodiment shown in fig. 2.
Referring to FIG. 2, one embodiment of predicting geochemical data for a sample dead zone in conjunction with sample acquisition and geographic factors includes the steps of:
s201, calculating the similarity between the sampling blind area and the sample collection point based on the geographic factors of the to-be-assigned area.
The geographic factors comprise geological factors and geospatial information, and the geological factors comprise geological backgrounds, soil types and land utilization types, so that the geospatial information, the geological backgrounds, the soil types and the land utilization types between the sampling blind areas and the sample collection points can be comprehensively calculated to obtain the similarity.
S202, judging whether the similarity exceeds a preset similarity threshold, if so, executing a step S203.
And if the similarity does not exceed the similarity threshold, not marking the corresponding sample acquisition point.
And S203, marking the corresponding sample collection point as a target sample collection point.
And S204, respectively configuring the assignment weights of all target sample acquisition points according to the similarity.
Wherein, a weight generation formula is constructed according to the similarity, and the formula is specifically as follows:
in the weight generation formula, w i Assigning a weight to the ith target sample acquisition Point, D i The similarity between the target sample collection points and the sampling blind areas is shown, and k is the number of the target sample collection points.
S205, the geochemical data of the sampling dead zone is predicted by combining the sample collection results of all the target sample collection points and the assigned weights.
The specific formula for predicting the geochemical data of the sampling dead zone is as follows:
wherein R is predicted geochemical data, w i Assigned weight, r, for the ith target sample acquisition point i Is the sample acquisition result of the ith target sample acquisition point.
The implementation principle of the embodiment is as follows:
because the distribution of the plurality of sample acquisition points is different, the difference between the geographic factors of the positions of the sample acquisition points and the geographic factors of the sampling blind areas is different, the similarity between each sample acquisition point and the sampling blind area is calculated firstly, then the sample acquisition points with high similarity are screened out through a preset similarity threshold value to be used as target sample acquisition points, different assignment weights are configured for all the target sample acquisition points according to the similarity, and finally the assignment weights of the target sample acquisition points and the sample acquisition results are combined for calculation, so that the geochemical data of the sampling blind areas are predicted.
In step S201 of the embodiment shown in fig. 2, geographic factor difference values between the sampling blind area and the sample collection point are calculated based on geographic factors, where the geographic factor difference values include differences of multiple geographic factors, different weight values are configured for different geographic factors, and finally, a similarity between the sampling blind area and the sample collection point is calculated. This is explained in detail with reference to the embodiment shown in fig. 3.
Referring to fig. 3, the calculation of the similarity between the sampling points of the sampling dead zone box based on the geographic factors includes the following steps:
s301, calculating a geographic factor difference value between the sampling blind area and the sample collection point based on geographic factors of the sampling blind area and the sample collection point.
The method comprises the steps of quantifying geographic spatial information, a geological background, a soil type and a land utilization type in geographic factors, and respectively calculating a geographic spatial information difference value, a geological background difference value, a soil type difference value and a land utilization type difference value between a quantified sampling blind area and a sample collection point.
S302, configuring an initial weight value for the geographic factor difference value.
And the initial weight value is a preset weight value.
And S303, optimizing the initial weight value through a preset group intelligent algorithm to obtain a final weight value.
The colony intelligent algorithm can adopt an ant colony intelligent algorithm.
And S304, calculating the similarity between the sampling blind area and the sample collection point by combining the final weight value and the geographic factor difference value.
The specific calculation formula of the similarity between the sampling blind area and the sample acquisition point is as follows:
wherein D is the similarity between the sampling blind area and the sample collection point, D 1 、d 2 、d 3 、d 4 Is divided into a geographic space information difference value, a geological background difference value, a soil type difference value and a land utilization type difference value, n 1 To n 4 And final weight values corresponding to the geographic space information difference value, the geological background difference value, the soil type difference value and the land utilization type difference value are respectively obtained.
The implementation principle of the embodiment is as follows:
when the similarity between the sampling blind area and the sample collection point is calculated, a plurality of geographic factors of the sampling blind area and the sample collection point need to be considered, the geographic factor difference values of the geographic factors can be calculated firstly, the initial weight values are configured for the geographic factor difference values, the initial weight values are optimized to be the final weight values through a group intelligent algorithm, the calculation of the similarity is facilitated to be more accurate, and the similarity between the sampling blind area and the sample collection point is calculated finally according to the geographic factor difference values and the corresponding final weight values.
In step S102 of the embodiment shown in fig. 1, another embodiment of this step may generate an adjacency relation between the sampling region and the sampling blind region by using geospatial information in the geographic factors, so as to predict geochemical data of a target sampling blind region adjacent to the sampling region by combining the adjacency relation and the sample acquisition result, and predict geochemical data of other sampling blind regions adjacent to the target sampling blind region according to the predicted geochemical data of the target sampling blind region. And the sampling area is a pattern spot where the sample acquisition result in the area to be assigned is located. This is explained in detail with reference to the embodiment shown in fig. 4.
Referring to FIG. 4, another embodiment of predicting geochemical data for a sample dead zone in conjunction with sample acquisition and geographic factors includes the steps of:
s401, generating adjacency relations between all sampling areas and all sampling blind areas based on the geographic space information of the areas to be assigned.
And generating adjacency relations between all sampling areas and all sampling blind areas according to the spatial position relations and the specific shapes of all sampling areas and all sampling blind areas in the area to be assigned.
S402, predicting geochemical data of a target sampling blind area adjacent to the sampling area based on the adjacency relation and according to the sample acquisition result.
The feature similarity between two adjacent areas can be calculated according to geographic factors, and the geochemical data of a target sampling blind area adjacent to the sampling area is predicted by combining the feature similarity and a sample acquisition result.
S403, judging whether the target sampling blind area has other adjacent sampling blind areas or not based on the adjacency relation, and if so, executing a step S404.
And if the target sampling blind area does not have other adjacent sampling blind areas, not executing other steps on the target sampling blind area.
S404, predicting the geochemical data of other sampling blind areas based on the geochemical data of the target sampling blind area.
The same method as that in step S402 can be used for the prediction method.
And S405, marking other sampling blind areas as target sampling blind areas.
And S406, repeating the judging step until all sampling blind areas in the area to be assigned are marked as target sampling blind areas.
And repeating the step S403 to the step S405, and when all the sampling dead zones are marked as target sampling dead zones, indicating that all the sampling dead zones in the to-be-assigned zone predict the geochemical data.
The implementation principle of the embodiment is as follows:
the method comprises the steps of firstly generating an adjacency relation between a sampling area and a sampling blind area according to the geospatial information of each area in an area to be assigned, wherein the adjacency relation between the two areas represents that the geospatial information between the two areas is similar, and the sampling area is an area for sampling and obtaining a sample acquisition result, so that the geochemical data of an adjacent target sampling blind area can be predicted according to the sample acquisition result of the sampling area, and after the geochemical data of the target sampling blind area is obtained through prediction, the geochemical data of other sampling blind areas adjacent to the target sampling blind area can be predicted continuously according to the geochemical data of the target sampling blind area, so that the data prediction of all sampling blind areas in the whole area to be assigned is completed through a data transmission prediction mode.
In step S402 of the embodiment shown in fig. 4, feature similarities between two adjacent regions are calculated according to geological factors, and a same sampling blind area may be adjacent to multiple sampling regions, so that it is necessary to predict the sampling blind area according to the number of sampling regions adjacent to the sampling blind area and the feature similarity between adjacent regions. This is explained in detail with reference to the embodiment shown in fig. 5.
Referring to fig. 5, predicting geochemical data for a target sample shadow adjacent a sample area based on adjacencies and sample acquisitions comprises the steps of:
and S501, calculating the feature similarity between all adjacent regions based on geological factors.
The calculation method may be the method used in the detailed description of step S304.
S502, judging whether a target sampling blind area adjacent to the sampling area is adjacent to other sampling areas, if not, executing a step S503; if yes, go to step S504.
S503, predicting geochemical data of the target sampling blind area according to the sample acquisition result of the sampling area.
S504, based on the feature similarity of the target sampling blind area and all adjacent sampling areas, different predicted value weights are configured for all the adjacent sampling areas respectively.
Wherein, the weight of the predicted value is a preset weight.
And S505, predicting the geochemical data of the target sampling blind area by combining the sample acquisition results and the predicted value weights of all adjacent sampling areas.
The implementation principle of the embodiment is as follows:
after generating an adjacency relation according to geospatial information, calculating the characteristic similarity of geological features between two adjacent regions according to geological factors of the two adjacent regions, before predicting data of target sampling blind regions adjacent to a sampling region, judging the number of the sampling regions adjacent to the target sampling blind region, and if the target sampling blind region is only adjacent to one sampling region, directly predicting the geochemical data of the target sampling blind region according to a sample acquisition result of the sampling region; if the target sampling blind area is adjacent to the plurality of sampling areas, different predicted value weights need to be configured for the plurality of adjacent sampling areas according to the feature similarity between the target sampling blind area and the plurality of adjacent sampling areas, the higher the feature similarity is, the larger the configured predicted value weight is, and finally, the geochemical data of the target sampling blind area is predicted by combining the predicted value weights corresponding to the adjacent sampling areas and the sample acquisition results.
In step S102 of the embodiment shown in fig. 1, another embodiment of this step may generate a virtual acquisition result through a preset deep learning model, continuously optimize the deep learning model according to the virtual acquisition result, and finally predict the geochemical data of the sampling blind area according to the virtual acquisition result of the virtual acquisition point located in the sampling blind area. This is explained in detail with reference to the embodiment shown in fig. 6.
Referring to FIG. 6, another embodiment of geochemical data for predicting sample dead zones by combining sample acquisition results and geographic factors of the area to be assigned comprises the steps of:
s601, fitting all sample acquisition results through a preset deep learning model for training to obtain a plurality of virtual sampling points and virtual acquisition results corresponding to the virtual sampling points.
The deep learning model may be a Cluster-GAN prediction model in a countermeasure network (GAN), and specifically, may be a generator in the Cluster-GAN prediction model, and the generator generates a virtual acquisition result by sampling in random discrete-continuous distribution.
S602, judging whether the virtual acquisition result is true or not based on the sample acquisition result, and if not, executing the step S603; if true, go to step S604.
The virtual acquisition result can be judged through a discriminator in the Cluster-GAN prediction model.
And S603, feeding back the judgment result to the deep learning model.
And S604, feeding back the judgment result to the deep learning model, and marking the target virtual sampling point corresponding to the virtual acquisition result in the region where the target virtual sampling point is located.
In this step, a discriminator described in step S602 may be additionally introduced, so as to strengthen the same-class constraint between the virtual acquisition result and the sample acquisition result.
And S605, counting the sampling number of the target virtual sampling points in the sampling blind area.
And S606, predicting geochemical data of the sampling dead zone by combining the virtual acquisition result of the target virtual sampling point and the sampling quantity.
The implementation principle of the embodiment is as follows:
the method comprises the steps of training to obtain a plurality of virtual acquisition results by taking an acquired actual sample acquisition result as a training basis of a deep learning model, judging the truth of the virtual acquisition result according to the sample acquisition result, and feeding back the judgment result to the deep learning model to optimize the deep learning model, so that the virtual acquisition result generated subsequently by the deep learning model can be closer to the sample acquisition result. If the judgment result is true, the target virtual sampling points corresponding to the virtual acquisition results can be used as sample sampling points to be marked to the corresponding positions in the area to be assigned, so that the geochemical data of the target sampling blind area can be predicted according to all the marked target virtual sampling points in the sampling blind area and the corresponding virtual acquisition results.
After step S601 in the embodiment shown in fig. 6, the virtual acquisition result may be reversely encoded and reversely input into the deep learning model to optimize the deep learning model. This is explained in detail with reference to the embodiment shown in fig. 7.
Referring to fig. 7, optimizing the deep learning model based on the virtual acquisition result includes the steps of:
and S701, reversely decoding all the virtual acquisition results to obtain a sample training set.
The reverse coding can be carried out through an encoder in a Cluster-GAN prediction model, and a sample training set is in discrete-continuous distribution.
S702, inputting the sample training set into the deep learning model.
Wherein the inverse decoded discrete-continuous distribution is inversely input to a generator in the Cluster-GAN prediction model.
The implementation principle of the embodiment is as follows:
the generated virtual acquisition result is reversely decoded into a sample training set, and the sample training set is input into the deep learning model, so that the training speed of the deep learning model can be accelerated, and the deep learning model can generate the virtual acquisition result close to the actual sample acquisition result of the sample more quickly.
In step S606 of the embodiment shown in fig. 6, the sampling number of the virtual sampling points in the sampling dead zone is determined according to the preset number threshold, and if the sampling number is insufficient, the virtual acquisition result in the sampling dead zone needs to be processed by an interpolation method, so as to predict the geochemical data of the sampling dead zone more accurately according to the virtual acquisition result. This is explained in detail with reference to the embodiment shown in fig. 8.
Referring to fig. 8, predicting geochemical data of a sampling blind area by combining a virtual acquisition result of a target virtual sampling point and the sampling number comprises the following steps:
s801, judging whether the sampling number exceeds a preset number threshold value, if so, executing a step S802; if not, step S803 is executed.
S802, geochemical data of the sampling dead zone is predicted by combining the virtual acquisition results of all the target virtual sampling points.
And S803, processing the virtual acquisition results of all the target virtual sampling points by an interpolation method to obtain a prediction curve.
And S804, predicting the geochemical data of the sampling dead zone according to the prediction curve.
The implementation principle of the embodiment is as follows:
if the number of the virtual sampling points in the sampling blind area is small, the prediction result is possibly inaccurate, so that the sampling number of the virtual sampling points in the sampling blind area is judged through a preset number threshold, and when the sampling number is less than the number threshold, the virtual acquisition results of all the virtual sampling points in the sampling blind area need to be completed through an interpolation method so as to generate a more accurate prediction curve to predict the geochemical data of the sampling blind area.
The embodiment of the application further discloses a assignment system for a geochemistry sampling dead zone, which includes a memory, a processor, and a program stored in the memory and running on the processor, and the program can be loaded and executed by the processor to implement the assignment method for a geochemistry sampling dead zone as shown in fig. 1 to 8.
The implementation principle of the embodiment is as follows:
through program calling, a sampling blind area which is not sampled is defined according to a sample acquisition result which is randomly sampled in the area to be assigned, and analysis on geological factors and geospatial information of the whole area to be assigned is added in the process of predicting the geochemical data of the sampling blind area according to the sample acquisition result, so that the predicted geochemical data is assigned to the sampling blind area which is not sampled.
An embodiment of the present application further discloses a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the computer program causes the processor to implement a method for assigning a geochemical sampling dead zone as shown in fig. 1 to 8.
The implementation principle of the embodiment is as follows:
through program calling, a sampling blind area which is not sampled is defined according to a sample acquisition result which is randomly sampled in the area to be assigned, and analysis on geological factors and geospatial information of the whole area to be assigned is added in the process of predicting the geochemical data of the sampling blind area according to the sample acquisition result, so that the predicted geochemical data is assigned to the sampling blind area which is not sampled.
The above embodiments are preferred embodiments of the present application, and the protection scope of the present application is not limited by the above embodiments, so: equivalent changes in structure, shape and principle of the present application shall be covered by the protection scope of the present application.
Claims (8)
1. A assignment method for geochemical sampling dead zones is characterized by comprising the following steps:
delineating a sampling blind area in the area to be assigned based on a sample acquisition result of the area to be assigned, wherein a pattern spot in which the sample acquisition result is located in the area to be assigned is a sampling area;
predicting the geochemical data of the sampling blind area by combining the sample acquisition result and the geographic factors of the area to be assigned, wherein the geographic factors comprise geological factors and geospatial information, and the geological factors comprise a geological background, a soil type and a land utilization type;
when the geospatial information is adopted to generate the adjacency relation between the sampling area and the sampling blind area, the step of predicting the geochemical data of the sampling blind area by combining the sample acquisition result and the geographic factors of the area to be assigned comprises the following steps:
generating adjacency relations between all sampling areas and all sampling blind areas based on the geospatial information of the area to be assigned;
calculating feature similarity between all adjacent regions based on the geological factors;
judging whether a target sampling blind area adjacent to the sampling area is adjacent to other sampling areas;
if the target sampling blind area is not adjacent to other sampling areas, predicting the geochemical data of the target sampling blind area according to the sample acquisition result of the sampling area;
if the target sampling blind area is adjacent to other sampling areas, configuring different predicted value weights for all adjacent sampling areas respectively based on the feature similarity of the target sampling blind area and all adjacent sampling areas;
predicting the geochemical data of the target sampling blind area by combining the sample acquisition results and the predicted value weights of all adjacent sampling areas;
judging whether the target sampling blind area has other adjacent sampling blind areas or not based on the adjacency relation;
if the target sampling blind area has other adjacent sampling blind areas, predicting the geochemical data of the other sampling blind areas based on the geochemical data of the target sampling blind area;
marking the other sampling blind areas as the target sampling blind areas;
repeating the judging step until all the sampling blind areas in the area to be assigned are marked as the target sampling blind areas;
assigning the geochemical data to the sampling blind area.
2. The assignment method for geochemical sampling dead zones according to claim 1, wherein when calculating the similarity between the sampling dead zone and a sample collection point according to the geographic factors of the area to be assigned, the predicting the geochemical data of the sampling dead zone by combining the sample collection result and the geographic factors of the area to be assigned comprises the following steps:
calculating the similarity between the sampling blind area and a sample acquisition point based on the geographic factors of the to-be-assigned area;
judging whether the similarity exceeds a preset similarity threshold value or not;
if the similarity exceeds the similarity threshold, marking the corresponding sample acquisition point as a target sample acquisition point;
respectively configuring the assignment weights of all target sample acquisition points according to the similarity;
and predicting the geochemical data of the sampling dead zone by combining the sample acquisition results of all target sample acquisition points and the assigned weights.
3. The assignment method for geochemical dead sampling areas according to claim 2, wherein the calculating the similarity between the dead sampling area and the sample collection point based on the geographic factors of the area to be assigned comprises the following steps:
calculating a geographic factor difference value between the sampling blind area and the sample collection point based on geographic factors of the sampling blind area and the sample collection point, wherein the geographic factor difference value comprises a geographic space information difference value, a geological background difference value, a soil type difference value and a land utilization type difference value;
configuring an initial weight value for the geographic factor difference value;
optimizing the initial weight value through a preset group intelligent algorithm to obtain a final weight value;
calculating the similarity between the sampling blind area and the sample collection point by combining the final weight value and the geographic factor difference value, wherein a specific calculation formula of the similarity between the sampling blind area and the sample collection point is as follows:
wherein D is the similarity between the sampling blind area and the sample collection point, D 1 、d 2 、d 3 、d 4 Is divided into a geographic space information difference value, a geological background difference value, a soil type difference value and a land utilization type difference value, n 1 To n 4 And final weight values corresponding to the geographic space information difference value, the geological background difference value, the soil type difference value and the land utilization type difference value are respectively obtained.
4. The assignment method for geochemical sampling dead zones according to claim 1, wherein the step of predicting the geochemical data of the sampling dead zone by combining the sample collection result and the geographic factor of the area to be assigned comprises the following steps:
fitting all the sample acquisition results through a preset deep learning model for training to obtain a plurality of virtual sampling points and virtual acquisition results corresponding to the virtual sampling points;
judging whether the virtual acquisition result is true or not based on the sample acquisition result;
if the virtual acquisition result is not true, feeding back a judgment result to the deep learning model;
if the virtual acquisition result is true, feeding back a judgment result to the deep learning model, and marking a target virtual sampling point corresponding to the virtual acquisition result in a located area;
counting the sampling number of the target virtual sampling points in the sampling blind area;
and predicting the geochemical data of the sampling blind area by combining the virtual acquisition result of the target virtual sampling point and the sampling quantity.
5. The assignment method for geochemical sampling blind areas according to claim 4, wherein the training is performed by fitting a preset deep learning model to all the sample collection results, and after obtaining a plurality of virtual sampling points and virtual collection results corresponding to the virtual sampling points, the method further comprises the following steps:
reversely decoding all the virtual acquisition results to obtain a sample training set;
inputting the sample training set into the deep learning model.
6. The method of assigning geochemical sampling dead zones according to claim 4, wherein said predicting the geochemical data of the sampling dead zone by combining the virtual acquisition result of the target virtual sampling point and the number of samples comprises the steps of:
judging whether the sampling number exceeds a preset number threshold value;
if the sampling quantity exceeds the quantity threshold value, predicting the geochemical data of the sampling blind area by combining the virtual acquisition results of all the target virtual sampling points;
if the sampling quantity does not exceed the quantity threshold, processing the virtual acquisition results of all target virtual sampling points by an interpolation method to obtain a prediction curve;
and predicting the geochemical data of the sampling blind area according to the prediction curve.
7. A assignment system for geochemical sampling dead zones, comprising a memory, a processor and a program stored on said memory and executable on said processor, said program being capable of being loaded and executed by the processor to implement a method of assigning geochemical sampling dead zones according to any one of claims 1 to 6.
8. A computer-readable storage medium, characterized in that it stores a computer program which, when executed by a processor, causes the processor to implement a method of assigning geosciences sample dead zones as claimed in any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210503858.4A CN114596498B (en) | 2022-05-10 | 2022-05-10 | Assignment method and system for geochemical sampling blind area and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210503858.4A CN114596498B (en) | 2022-05-10 | 2022-05-10 | Assignment method and system for geochemical sampling blind area and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114596498A CN114596498A (en) | 2022-06-07 |
| CN114596498B true CN114596498B (en) | 2022-07-29 |
Family
ID=81821872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210503858.4A Active CN114596498B (en) | 2022-05-10 | 2022-05-10 | Assignment method and system for geochemical sampling blind area and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114596498B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103745499A (en) * | 2013-12-27 | 2014-04-23 | 中国石油天然气股份有限公司 | Method for field geological modeling based on public geographic information image data |
| CN112559648A (en) * | 2019-09-10 | 2021-03-26 | 广州市地质调查院(广州市地质环境监测中心) | Intelligent geochemical survey sampling point layout method |
| CN114139819A (en) * | 2021-12-07 | 2022-03-04 | 成都理工大学 | Geochemical variable space prediction method based on geostatistical weighted random forest |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9158016B2 (en) * | 2012-04-30 | 2015-10-13 | Conocophillips Company | Multi-dimensional data reconstruction constrained by a regularly interpolated model |
| US10267937B2 (en) * | 2014-04-17 | 2019-04-23 | Saudi Arabian Oil Company | Generating subterranean imaging data based on vertical seismic profile data and ocean bottom sensor data |
| CN104899429B (en) * | 2015-05-14 | 2017-07-28 | 湖北省地质调查院 | It is a kind of based on multifactor geochemical elements content assignment method and system |
| CN109541172B (en) * | 2018-10-25 | 2019-12-17 | 北京农业信息技术研究中心 | Soil attribute value calculation method and device |
| CN111046120B (en) * | 2019-11-28 | 2022-08-19 | 安徽农业大学 | Extreme air temperature semantic inverse distance weight interpolation method |
| CN114387531A (en) * | 2022-01-19 | 2022-04-22 | 重庆邮电大学 | Ground surface temperature downscaling method based on improved geographic weighted regression model |
| CN114154627B (en) * | 2022-02-10 | 2022-05-20 | 山东省地质矿产勘查开发局第七地质大队（山东省第七地质矿产勘查院） | Soil profile measuring method and device based on GIS and double-layer neural network |
-
2022
- 2022-05-10 CN CN202210503858.4A patent/CN114596498B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103745499A (en) * | 2013-12-27 | 2014-04-23 | 中国石油天然气股份有限公司 | Method for field geological modeling based on public geographic information image data |
| CN112559648A (en) * | 2019-09-10 | 2021-03-26 | 广州市地质调查院(广州市地质环境监测中心) | Intelligent geochemical survey sampling point layout method |
| CN114139819A (en) * | 2021-12-07 | 2022-03-04 | 成都理工大学 | Geochemical variable space prediction method based on geostatistical weighted random forest |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114596498A (en) | 2022-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109829399B (en) | Vehicle-mounted road scene point cloud automatic classification method based on deep learning | |
| Vollering et al. | Bunching up the background betters bias in species distribution models | |
| CN111199214B (en) | Residual network multispectral image ground object classification method | |
| Brus et al. | Sampling for validation of digital soil maps | |
| US8682624B2 (en) | Deterministic version of the multiple point geostatistics simulation/reconstruction method with the simulated/reconstructed values are directly taken from the training images without prior estimation of the conditional | |
| AU2013274839B2 (en) | System and method for optimizing the number of conditioning data in multiple point statistics simulation | |
| CN106919957B (en) | Method and device for processing data | |
| GB2547816A (en) | Actually-measured marine environment data assimilation method based on sequence recursive spare filtering three-dimensional variation | |
| CN104899849A (en) | Multi-target interactive image segmentation method and device | |
| US11687804B2 (en) | Latent feature dimensionality bounds for robust machine learning on high dimensional datasets | |
| CN115145906A (en) | Preprocessing and completion method for structured data | |
| CN105405118A (en) | Underwater sonar image target detection method based on hybrid quantum derivative frog leaping | |
| WO2014072810A1 (en) | Method for processing an image | |
| CN115457492A (en) | Target detection method and device, computer equipment and storage medium | |
| Broennimann et al. | Package ‘ecospat’ | |
| CN114792044A (en) | Intelligent early warning method and system for settlement of foundation pit adjacent to earth surface by coupling spatial characteristics | |
| CN110490061B (en) | Uncertainty modeling and measuring method for remote sensing image characteristics | |
| CN109961129A (en) | A kind of Ocean stationary targets search scheme generation method based on improvement population | |
| CN114596498B (en) | Assignment method and system for geochemical sampling blind area and storage medium | |
| CN104573339A (en) | Method and device for determining geological parameters of shale gas reservoir | |
| Blackford et al. | Incorporating spatial uncertainty maps into soil sampling improves digital soil mapping classification accuracy in Ontario, Canada | |
| CN115292971B (en) | Bayes-based crack attribute analysis method and device and storage medium | |
| Packianather et al. | Modelling neural network performance through response surface methodology for classifying wood veneer defects | |
| CN114758123A (en) | Remote sensing image target sample enhancement method | |
| CN111767980B (en) | Model optimization method, device and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |