Data acquisition and analysis system for river basin ecological product monitoring

Abstract

The invention relates to the technical field of data processing, and discloses a data acquisition and analysis system for river basin ecological product monitoring, and the system comprises the steps: collecting the monitoring data of M ecological products in a target river basin according to a fixed time interval t in a confidence time period T, and the amplitude data of the corresponding monitoring data; wherein the monitoring data represents the dimension value of the ecological product of the unit dimension; constructing a first topological network based on the M ecological products of the target drainage basin, and constructing a first verification model according to the first topological network; dividing the target drainage basin into K sub-regions from upstream to downstream according to a fixed distance interval so as to construct a second topology network, and constructing a second verification model according to the second topology network; and in the target time period, obtaining monitoring data of the M ecological products of the target drainage basin, and performing first verification and second verification on the monitoring data through the first verification model and the second verification model to obtain a human interference result of the target drainage basin.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically, it relates to a data acquisition and analysis system for monitoring watershed ecological products. Background Art

[0002] With the increasing emphasis on environmental protection and ecological civilization construction at present, the protection of the watershed ecological environment has gradually shifted from simple pollution control to more proactive ecological conservation and restoration. Compared with the past, at the present stage, more emphasis is placed on the self-recovery ability of the ecosystem, and human intervention is minimized as much as possible.

[0003] The existing watershed ecological environment measures the data of ecological products in the target watershed at fixed time intervals to analyze the value of the watershed ecosystem in evaluation systems such as GEP. However, the existing analysis methods of monitoring data are difficult to determine whether there is human intervention in the watershed ecological environment. If there is human intervention in the watershed ecological environment, the GEP evaluation result of the target watershed will be inaccurate.

[0004] Therefore, there is an urgent need for a system for confidence analysis and judgment of human intervention based on monitoring data. Summary of the Invention

[0005] The present invention provides a data acquisition and analysis system for monitoring watershed ecological products to solve the technical problems raised in the background art.

[0006] The present invention provides a data acquisition and analysis system for monitoring watershed ecological products, including:

[0007] A data acquisition module, configured to collect monitoring data of M ecological products in the target watershed and the amplitude data of the corresponding monitoring data at a fixed time interval t within a confidence time period T; wherein, the monitoring data represents the dimensional value of the ecological product with a unit dimension;

[0008] A first processing module, configured to construct a first topological network based on M ecological products in the target watershed and construct a first verification model according to the first topological network;

[0009] A second processing module, configured to divide the target watershed into K sub-regions at a fixed distance interval from upstream to downstream to construct a second topological network and construct a second verification model according to the second topological network;

[0010] An interference analysis module, configured to obtain the monitoring data of M ecological products in the target watershed within a target time period and perform a first verification and a second verification on the monitoring data through the first verification model and the second verification model respectively to obtain the human intervention result of the target watershed.

[0011] Further, the confidence time period indicates that within a pre-determined time period, there is no human interference in the monitoring data collected for several types of ecological products in the target basin at fixed time intervals.

[0012] Further, the amplitude data includes: obtaining the monitoring data of the m-th type of ecological product at adjacent moments in the target basin within the confidence time period, and calculating the difference between the corresponding monitoring data to obtain the amplitude data of the corresponding monitoring data; where 1 ≤ k ≤ K, 1 ≤ m ≤ M, and both m and k are positive integers.

[0013] Further, constructing a topological network based on the M types of ecological products in the target basin includes: classifying the M types of ecological products in the target basin;

[0014] The feature types include: if the m-th type of ecological product belongs to the biological category, it is classified as the first type; if the m-th type of ecological product belongs to the environmental category, it is classified as the second type; if the m-th type of ecological product belongs to the GEP system index category, it is classified as the third type;

[0015] Mapping the ecological products of the first type to the first topological nodes respectively, and if there is a predation relationship or a symbiotic relationship between the ecological products corresponding to any two first topological nodes, a topological edge is constructed between the corresponding first topological nodes;

[0016] Mapping the ecological products of the second type to the second topological nodes respectively, and constructing a topological edge between each second topological node and each first topological node;

[0017] Mapping the ecological products of the third type to the third topological nodes respectively, and constructing a topological edge between each third topological node and each first topological node and second topological node to obtain the first topological network.

[0018] Further, constructing a first verification model based on the first topological network includes:

[0019] Obtaining the amplitude data of the M types of ecological products at each adjacent moment within the confidence time period, and constructing a first verification model based on the B network, specifically including:

[0020] Taking the amplitude data of the ecological products corresponding to the topological nodes that have topological edges with the topological node corresponding to the m-th type of ecological product in the first topological network at the i-th moment as the first initial sample data; taking the amplitude data of the m-th type of ecological product at the i-th moment as the first initial sample label, where, and i is a positive integer;

[0021] After normalizing both the first initial sample data and the first initial sample labels, the first sample data and the first sample labels are obtained; and a first verification model is trained based on the first sample data and the first sample labels; wherein, the first verification model is constructed based on a one-dimensional convolutional neural network and converges through a mean squared error loss function.

[0022] Further, a second topological network is constructed, including:

[0023] Each sub-region is mapped to a fourth topological node, and topological edges are constructed between the fourth topological nodes corresponding to adjacent sub-regions to obtain the second topological network.

[0024] Further, a second verification model is constructed, including:

[0025] Identify the common ecological products in the K sub-regions; wherein, the common ecological products refer to the ecological products that exist in all K sub-regions, and the monitoring data of each ecological product is greater than or equal to a preset monitoring threshold at any moment within the confidence time period;

[0026] Vectorize the amplitude data of the common ecological products at the i-th moment in the sub-regions corresponding to the topological nodes that have topological edges with the topological nodes corresponding to the k-th sub-region in the second topological network to obtain the second initial sample data; vectorize the amplitude data of the common ecological products in the k-th sub-region at the i-th moment to obtain the second initial sample labels;

[0027] After normalizing the second initial sample data and the second initial sample labels, the second sample data and the second sample labels are obtained; and a second verification model is trained based on the second sample data and the second sample labels; wherein, the second verification model is constructed based on a one-dimensional convolutional neural network and converges through a mean squared error loss function.

[0028] Further, the first verification includes:

[0029] During the target time period T gl After normalizing the amplitude data of the m-th ecological product at the j-th moment, it is used as the first data to be matched; And j is a positive integer;

[0030] In the first topological network, the amplitude data of the ecological products corresponding to the topological nodes that have topological edges with the topological node corresponding to the m-th ecological product is used as the first input data;

[0031] Normalize the first input data and then input it into the first verification model to obtain the output data of the first verification model; if the difference between the output data of the first verification model and the first data to be matched is less than or equal to a preset first matching threshold, the matching is successful and the first verification is obtained, otherwise the matching fails.

[0032] Further, the second verification includes:

[0033] During the target time period, normalize the amplitude data of the common ecological products in the k-th sub-region and use it as the second data to be matched;

[0034] In the second topological network, obtain the amplitude data of the sub-region corresponding to the topological node that has a topological edge with the topological node corresponding to the k-th sub-region as the second input data;

[0035] Normalize the second input data and then input it into the second verification model to obtain the output data of the second verification model; if the Euclidean distance between the output data of the second verification model and the second data to be matched is less than or equal to a preset second matching threshold, the matching is successful and the second verification is obtained, otherwise the matching fails.

[0036] Further, the human interference results in the target basin include:

[0037] During the target time period, if the amplitude data of any ecological product in the target basin at any moment passes the first verification, and the amplitude data of the common ecological products in any sub-region of the target basin at any moment passes the second verification, then there is no human interference in the target basin, otherwise there is human interference in the target basin.

[0038] The beneficial effects of the present invention are as follows: By constructing a dual verification model, using the internal data association between ecological products to construct the first verification model, and at the same time dividing the basin into multiple sub-regions according to the data change direction to construct the second verification model, it effectively distinguishes the monitoring data changes caused by natural ecological changes and human interference, thereby greatly improving the credibility of the data used for the GEP assessment of the basin ecosystem. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a module diagram of a data acquisition and analysis system for monitoring ecological products in a basin according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0040] Reference will now be made to example embodiments to discuss the subject matter described herein. It should be understood that discussing these embodiments is only to enable those skilled in the art to better understand and thus implement the subject matter described herein. Without departing from the scope of protection of the content of this specification, changes can be made to the functions and arrangements of the elements discussed. Each example can omit, substitute, or add various processes or components as needed. Additionally, features described relative to some examples can also be combined in other examples.

[0041] As Figure 1 shown, a data acquisition and analysis system for monitoring ecological products in a river basin includes:

[0042] A data acquisition module for collecting monitoring data of M ecological products in the target river basin and the amplitude data of the corresponding monitoring data at a fixed time interval t within a confidence time period T; wherein, the monitoring data represents the dimensional value of the ecological product with a unit dimension.

[0043] A first processing module for constructing a first topological network based on M ecological products in the target river basin and constructing a first verification model according to the first topological network.

[0044] A second processing module for dividing the target river basin from upstream to downstream into K sub-regions at a fixed distance interval to construct a second topological network and constructing a second verification model according to the second topological network.

[0045] An interference analysis module for obtaining the monitoring data of M ecological products in the target river basin within the target time period and respectively performing a first verification and a second verification on the monitoring data through the first verification model and the second verification model to obtain the human interference result of the target river basin.

[0046] It should be noted that "the monitoring data represents the dimensional value of the ecological product with a unit dimension" means that a unified standardization or normalization measurement method is adopted for various ecological products in the target river basin to establish a comparable numerical system. The "unit dimension" can be determined according to the characteristics of the ecological product, the monitoring target, and industry standards. For example, aquatic organisms can be represented by the number of individuals, biomass, or density per cubic meter or per hectare, vegetation can be represented by the coverage rate or biomass per unit area, and water quality parameters are represented by concentration or percentage. This method ensures that different ecological product data is convenient for management, comparison, and comprehensive analysis under unified measurement, thereby improving the accuracy of identifying the overall state of the ecosystem and potential human interference. Whether the data is numerical, proportional, or concentration-based, as long as it is expressed in a unified dimension, it can meet the requirements of this application.

[0047] In an embodiment of the present invention, the confidence time period means that within a pre-determined time period, the monitoring data collected for several ecological products in the target river basin at a fixed time interval is free from human interference.

[0048] Specifically, within the confidence time period, the collected data is considered to truly reflect the natural ecological state of the basin without external human interference, thus providing a reliable basis for constructing a verification model based on this data later. Human interference includes: artificially releasing ecological products in the target basin to obtain a false value assessment of GEP and other behaviors.

[0049] In an embodiment of the present invention, the amplitude data includes: obtaining the monitoring data of the m-th type of ecological product at adjacent moments in the target basin within the confidence time period, and calculating the difference between the corresponding monitoring data to obtain the amplitude data of the corresponding monitoring data; where 1 ≤ k ≤ K, 1 ≤ m ≤ M, and both m and k are positive integers.

[0050] Specifically, the amplitude data is used to quantitatively describe the degree of change of ecological products in a short time. For example, if the monitoring data of the m-th type of ecological product at the i-th moment is 10, and the monitoring data of the m-th type of ecological product at the (i - 1)-th moment is 13, then the amplitude data at the i-th moment is -3.

[0051] In an embodiment of the present invention, constructing a topological network based on M types of ecological products in the target basin includes: classifying the M types of ecological products in the target basin;

[0052] The feature types include: if the m-th type of ecological product belongs to the biological category, it is classified into the first type; if the m-th type of ecological product belongs to the environmental category, it is classified into the second type; if the m-th type of ecological product belongs to the GEP system index category, it is classified into the third type;

[0053] Mapping the ecological products of the first type to the first topological nodes respectively, and if there is a predation relationship or a symbiotic relationship between the ecological products corresponding to any two first topological nodes, a topological edge is constructed between the corresponding first topological nodes;

[0054] Mapping the ecological products of the second type to the second topological nodes respectively, and constructing a topological edge between each second topological node and each first topological node;

[0055] Mapping the ecological products of the third type to the third topological nodes respectively, and constructing a topological edge between each third topological node and each first topological node and second topological node to obtain the first topological network.

[0056] It should be noted that the first topological network is used to reflect the internal dependence relationship between various types of ecological products in the target basin and the transmission mechanism of data changes. By classifying ecological products into biological, inorganic, and evaluation index categories, the ecological products are thus classified into three categories.

[0057] For biological ecological products, the topology is established based on the predator-prey relationship and the symbiotic relationship. The predator-prey relationship refers to the ecological interaction in which one organism captures, ingests and digests another organism to obtain energy and nutrients. The symbiotic relationship refers to the long-term and close interaction between different species of organisms, in which at least one of them benefits.

[0058] For example, there are four ecological products, namely, algae 1, algae 2, herbivorous fish and carnivorous fish. Among them, algae 1 and algae 2 are in a symbiotic relationship. Herbivorous fish eat algae 1 and algae 2, so herbivorous fish establish a predator-prey relationship with algae 1 and algae 2 respectively. Carnivorous fish eat herbivorous fish, so carnivorous fish and herbivorous fish establish a predator-prey relationship. Correspondingly, when the density per unit area of ​​algae 1 (monitoring data) changes (amplitude data), the monitoring data of algae 2, herbivorous fish and carnivorous fish will also change to varying degrees (amplitude data).

[0059] For environmental ecological products, they provide living space for biological ecological products. Therefore, when the monitoring data of each environmental ecological product changes (amplitude data), the monitoring data of biological ecological products also changes (amplitude data).

[0060] For the ecological products of evaluation index type, it is a concept defined by humans. Therefore, when the monitoring data of each biological ecological product and environmental ecological product changes (amplitude data), the monitoring data of the corresponding ecological products of evaluation index type also changes.

[0061] In one embodiment of the present invention, constructing a first verification model based on a first topology network includes:

[0062] Obtain the amplitude data of M ecological products at each adjacent moment within the confidence period, and build the first verification model based on the B network, including:

[0063] The amplitude data of the ecological product corresponding to the topological node with a topological edge established with the topological node corresponding to the mth ecological product in the first topological network at the i-th moment is used as the first initial sample data; the amplitude data of the mth ecological product at the i-th moment is used as the first initial sample label, where, And i is a positive integer;

[0064] After normalizing the first initial sample data and the first initial sample label, the first sample data and the first sample label are obtained; and a first verification model is obtained by training based on the first sample data and the first sample label; wherein the first verification model is constructed based on a one-dimensional convolutional neural network and converges through a mean square error loss function.

[0065] In one embodiment of the present invention, both the first verification model and the second verification model are constructed based on the one-dimensional convolutional neural network structure in deep learning. This structure effectively improves the model's learning ability for non-linear and heterogeneous data by extracting feature patterns in ecological product monitoring data layer by layer.

[0066] Deep learning models usually consist of multiple convolutional layers, activation layers (such as ReLU), pooling layers, and fully connected layers. In the present invention, the amplitude data of ecological products associated with each topological node in the first verification model is used as the input sequence, and feature extraction is performed through multiple convolutional kernels to obtain high-dimensional deep representation features. This process uses a sliding window mechanism to scan time series data and extract its local change trend to realize the modeling of the dynamic features of ecological products.

[0067] During the training process, the system uses the mean square error (MSE) loss function as the optimization objective based on the sample data and labels collected within the confidence time period, and uses the gradient descent algorithm to iteratively update the model parameters until the loss function converges. The entire training process adopts an end-to-end supervised learning method to ensure that the model has high accuracy in the prediction task of the target ecological product.

[0068] The model also includes batch normalization, Dropout layers to prevent overfitting, and an Early Stopping mechanism during training, thereby improving the generalization performance and avoiding underfitting or overfitting phenomena on small sample or high-volatility data sets.

[0069] It should be noted that the first topological network integrates the supporting effects of biological ecological products and environmental ecological products on biology, as well as the associations between evaluation index ecological products and the first two types of data. By obtaining the amplitude data of the m-th ecological product and the amplitude data of the ecological products associated with it, sample data and sample labels are constructed, and then the first verification model is trained.

[0070] In one embodiment of the present invention, the first verification model includes M hidden units. Among them, the m-th hidden unit is used to input the normalized vector constructed from the amplitude data of the ecological products corresponding to the topological nodes that have an edge relationship with the topological node corresponding to the m-th ecological product, and output the predicted amplitude data of the m-th ecological product.

[0071] In one embodiment of the present invention, a second topological network is constructed, including:

[0072] Each sub-region is mapped to a fourth topological node, and topological edges are constructed between the fourth topological nodes corresponding to adjacent sub-regions to obtain the second topological network.

[0073] It should be noted that the second topological network is used to reflect the spatial diffusion and change relationship of ecological product monitoring data in the basin from upstream to downstream. Specifically, the basin is divided into several sub-regions at a fixed distance, each sub-region is mapped to a fourth topological node, and topological edges are constructed between adjacent sub-region nodes, thus forming the second topological network. The second topological network can capture how the impact gradually spreads along the basin to the downstream when the monitoring data of any ecological product in the upstream changes, and then show the trend of data changes in the whole basin.

[0074] In an embodiment of the present invention, constructing a second verification model includes:

[0075] Determine the common ecological products in K sub-regions; wherein, the common ecological products refer to the ecological products that exist in all K sub-regions, and the monitoring data of each ecological product is greater than or equal to the preset monitoring threshold at any moment within the confidence time period;

[0076] Vectorize the amplitude data of the common ecological products in the sub-region corresponding to the topological node with which the topological node corresponding to the k-th sub-region in the second topological network has a topological edge at the i-th moment to obtain the second initial sample data; vectorize the amplitude data of the common ecological products in the k-th sub-region at the i-th moment to obtain the second initial sample label;

[0077] After normalizing the second initial sample data and the second initial sample label, obtain the second sample data and the second sample label; and train a second verification model based on the second sample data and the second sample label; wherein, the second verification model is constructed based on a one-dimensional convolutional neural network and converges through a mean square error loss function.

[0078] Specifically, in the second verification model, since the input data comes from the sub-regions after spatial division, its amplitude data also extracts temporal features through a one-dimensional convolutional neural network, and finally the Euclidean distance between the generated prediction vector and the actual amplitude vector is calculated as the discrimination criterion. In this process, the number of layers and channels of the deep neural network are adjusted according to the complexity of the basin and the data dimension to ensure that the model has sufficient expressive ability. Through the above process of constructing a one-dimensional convolutional neural network based on deep learning, the present invention effectively improves the model's understanding ability of the internal structure and spatial propagation law of ecological monitoring data, thereby enhancing the robustness and reliability of ecological interference discrimination on the premise of ensuring the system structure remains unchanged.

[0079] It should be noted that on the basis of dividing the basin into several sub-regions at fixed intervals from top to bottom, the second topological network is used to reflect the changing trend of the amplitude data of the common ecological products in each sub-region. Among them, the common ecological products refer to the ecological products whose monitoring data is greater than or equal to the threshold at each moment within the confidence time period. For example, within the confidence time period, if the monitoring data of algae A at each moment in K sub-regions is greater than the threshold of 100, then algae A is regarded as a common ecological product.

[0080] Specifically, the system first determines the common ecological products that exist in all K sub-regions, then calculates the amplitude data of the common ecological products collected in each sub-region within the confidence time period, and vectorizes these amplitude data to form the second initial sample data; at the same time, the vectorized amplitude data of the common ecological products in each sub-region at the corresponding moment is used as the second initial sample label. After normalization, a one-dimensional convolutional neural network and a mean squared error loss function are used to train the second initial sample data and labels, so as to obtain the second verification model. This model can effectively reflect the spatial diffusion and changing trend of the common ecological product data in the basin from upstream to downstream, and then judge whether the data change conforms to the natural propagation law.

[0081] In an embodiment of the present invention, the second verification model includes K hidden units, where the k-th hidden unit is used to input the normalized vector constructed from the amplitude data of the common ecological products in the sub-region corresponding to the topological node that has an edge relationship with the topological node corresponding to the k-th sub-region, and output the predicted amplitude data of the common ecological products in the k-th sub-region.

[0082] In an embodiment of the present invention, the first verification includes:

[0083] Within the target time period T gl After normalizing the amplitude data of the m-th ecological product at the j-th moment, it is used as the first data to be matched; And j is a positive integer;

[0084] In the first topological network, the amplitude data of the ecological products corresponding to the topological nodes that have topological edges with the topological node corresponding to the m-th ecological product is used as the first input data;

[0085] After normalizing the first input data, it is input into the first verification model to obtain the output data of the first verification model; if the difference between the output data of the first verification model and the first data to be matched is less than or equal to the preset first matching threshold, the matching is successful and the first verification is obtained, otherwise the matching fails.

[0086] It should be noted that if the difference between the predicted value output by the first verification model and the actual monitored value exceeds the preset threshold, it can be determined that there is potential human interference or abnormal change; on the contrary, if the difference is within a reasonable range, it indicates that the change of the ecological product at this moment conforms to the prediction of the system for the natural ecological relationship, thus completing the first verification process.

[0087] In an embodiment of the present invention, the second verification includes:

[0088] During the target time period, after normalizing the amplitude data of the common ecological products in the k-th sub-region, it is used as the second data to be matched;

[0089] In the second topological network, obtain the amplitude data of the sub-region corresponding to the topological node that has a topological edge established with the topological node corresponding to the k-th sub-region as the second input data;

[0090] After normalizing the second input data, input it into the second verification model to obtain the output data of the second verification model; if the Euclidean distance between the output data of the second verification model and the second data to be matched is less than or equal to the preset second matching threshold, the matching is successful and the second verification is obtained, otherwise the matching fails.

[0091] It should be noted that if the difference (Euclidean distance) between the model output result and the actual monitored data of the k-th sub-region does not exceed the set threshold, it indicates that the change trend of the monitored data in this sub-region is consistent with the natural diffusion law of the basin, and the second verification is passed; if the difference exceeds the threshold, it may mean that there are abnormal fluctuations or human interference in this area.

[0092] In an embodiment of the present invention, the human interference result of the target basin includes:

[0093] During the target time period, if the amplitude data of any ecological product in the target basin at any moment all pass the first verification, and the amplitude data of the common ecological products in any sub-region of the target basin at any moment all pass the second verification, then there is no human interference in the target basin, otherwise there is human interference in the target basin.

[0094] Specifically, the determination logic of human interference in the target basin can be summarized as follows: Passing the first verification indicates that within the target time period, the amplitude data of any ecological product at any moment conforms to the reasonable change range predicted based on the first topological network (the internal relationship of ecological products); passing the second verification indicates that within the target time period, the amplitude data of the common ecological products in any sub-region at any moment conforms to the reasonable change range predicted based on the second topological network (the spatial diffusion relationship from upstream to downstream); and only when both verifications are passed, the system determines that there is no human interference in the target basin; if any one of the verifications fails, it is inferred that there is human interference in the target basin.

[0095] The above describes the embodiments of this example, but this example is not limited to the above specific implementation manners. The above specific implementation manners are merely illustrative rather than restrictive. Under the inspiration of this example, those of ordinary skill in the art can also make many forms, all of which fall within the protection scope of this example.

Claims

1. A data collection and analysis system for watershed ecological product monitoring, characterized in that: include: The data collection module is used to collect monitoring data of M kinds of ecological products in the target watershed at fixed time intervals t within the confidence period T, as well as the amplitude data of the corresponding monitoring data; wherein the monitoring data represents the dimension value of the ecological product in unit dimension; A first processing module is used to construct a first topological network based on the M ecological products of the target watershed, and to construct a first verification model according to the first topological network; The second processing module is used to divide the target watershed from upstream to downstream into K sub-areas at fixed distance intervals to construct a second topological network, and to construct a second verification model based on the second topological network; The interference analysis module is used to obtain the monitoring data of M kinds of ecological products in the target watershed within the target time period, and perform the first verification and the second verification on the monitoring data through the first verification model and the second verification model respectively to obtain the human interference result of the target watershed.

2. A data collection and analysis system for watershed ecological product monitoring according to claim 1, characterized in that: The confidence period means that within a predetermined period of time, there is no human interference in the monitoring data collected on several ecological products in the target watershed at fixed time intervals.

3. A data collection and analysis system for watershed ecological product monitoring according to claim 2, characterized in that: Amplitude data, including: obtaining the monitoring data of the mth ecological product at adjacent moments in the target watershed within the confidence period, calculating the difference between the corresponding monitoring data to obtain the amplitude data of the corresponding monitoring data; wherein, 1≤k≤K, 1≤m≤M, m and k are both positive integers.

4. A data collection and analysis system for watershed ecological product monitoring according to claim 3, characterized in that: A topological network is constructed based on the M kinds of ecological products in the target watershed, including: classifying the M kinds of ecological products in the target watershed; The characteristic types include: if the mth ecological product belongs to the biological category, it is classified as the first type; if the mth ecological product belongs to the environmental category, it is classified as the second type; if the mth ecological product belongs to the GEP system indicator category, it is classified as the third type; The first type of ecological products are respectively mapped to first topological nodes. If any two ecological products corresponding to the first topological nodes have a predator-prey relationship or a symbiotic relationship, a topological edge is constructed between the corresponding first topological nodes. Mapping the second type of ecological products to second topological nodes respectively, and constructing a topological edge between each second topological node and each first topological node; The third type of ecological products are respectively mapped to third topological nodes, and topological edges are constructed between each third topological node and each first topological node and second topological node to obtain a first topological network.

5. A data collection and analysis system for watershed ecological product monitoring according to claim 4, characterized in that: Building a first verification model based on the first topology network includes: Obtain the amplitude data of M ecological products at each adjacent moment within the confidence period, and build the first verification model based on the B network, including: The amplitude data of the ecological product corresponding to the topological node with a topological edge established with the topological node corresponding to the mth ecological product in the first topological network at the i-th moment is used as the first initial sample data; the amplitude data of the mth ecological product at the i-th moment is used as the first initial sample label, where, And i is a positive integer; After normalizing the first initial sample data and the first initial sample label, the first sample data and the first sample label are obtained; and a first verification model is obtained by training based on the first sample data and the first sample label; wherein the first verification model is constructed based on a one-dimensional convolutional neural network and converges through a mean square error loss function.

6. A data collection and analysis system for watershed ecological product monitoring according to claim 5, characterized in that: Construct the second topology network, including: Each sub-region is mapped to a fourth topological node, and topological edges are constructed between the fourth topological nodes corresponding to adjacent sub-regions to obtain a second topological network.

7. A data collection and analysis system for watershed ecological product monitoring according to claim 6, characterized in that: Construct the second validation model, including: Determine the common ecological products in the K sub-areas; wherein the common ecological products refer to ecological products that exist in the K sub-areas, and the monitoring data of each ecological product is greater than or equal to the preset monitoring threshold at any time within the confidence period; Vectorize the amplitude data of the shared ecological products in the sub-region corresponding to the topological node that has a topological edge with the topological node corresponding to the k-th sub-region in the second topological network at the i-th moment to obtain the second initial sample data; vectorize the amplitude data of the shared ecological products in the k-th sub-region at the i-th moment to obtain the second initial sample label; After normalizing the second initial sample data and the second initial sample label, the second sample data and the second sample label are obtained; and the second verification model is trained based on the second sample data and the second sample label; wherein the second verification model is constructed based on a one-dimensional convolutional neural network and converges through a mean square error loss function.

8. A data collection and analysis system for watershed ecological product monitoring according to claim 7, characterized in that: The first verification includes: In the target time period T gl Within, the amplitude data of the mth ecological product at the jth moment is normalized and used as the first data to be matched; And j is a positive integer; In the first topological network, the amplitude data of the ecological product corresponding to the topological node with the topological edge established with the topological node corresponding to the mth ecological product is used as the first input data; The first input data is normalized and input into the first verification model to obtain the output data of the first verification model; if the difference between the output data of the first verification model and the first data to be matched is less than or equal to the preset first matching threshold, the match is successful and the first verification is obtained, otherwise the match fails.

9. A data collection and analysis system for watershed ecological product monitoring according to claim 8, characterized in that: The second verification includes: In the target time period, the amplitude data of the common ecological products in the kth sub-region are normalized and used as the second data to be matched; In the second topological network, the amplitude data of the sub-region corresponding to the topological node having the topological edge established with the topological node corresponding to the k-th sub-region is obtained as the second input data; The second input data is normalized and input into the second verification model to obtain the output data of the second verification model; if the Euclidean distance between the output data of the second verification model and the second data to be matched is less than or equal to the preset second matching threshold, the match is successful and the second verification is obtained, otherwise the match fails.

10. A data collection and analysis system for watershed ecological product monitoring according to claim 9, characterized in that: The results of human intervention in the target watershed, including: During the target time period, if the amplitude data of any ecological product in the target basin at any time passes the first verification, and the amplitude data of the common ecological product in any sub-area in the target basin at any time passes the second verification, then there is no human interference in the target basin; otherwise, there is human interference in the target basin.

Images