Biodiversity protection area monitoring system

Through the combination of intelligent sensor networks and drones, the multimodal deep learning model is used to monitor ecological events, which solves the problem that the dynamic changes of ecosystems cannot be evaluated in real time in the existing technology, and improves the monitoring efficiency and protection of biodiversity protected areas. ability.

CN119046881AInactive Publication Date: 2024-11-29NANJING INST OF ENVIRONMENTAL SCI MINIST OF ECOLOGY & ENVIRONMENT OF THE PEOPLES REPUBLIC OF CHINA

Patent Information

Application Number
CN202411164100.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-23
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing biodiversity protected area monitoring system cannot accurately evaluate the dynamic changes in the ecosystem in real time, resulting in poor monitoring effects.

Method used

Using a combination of intelligent sensor networks and drones, the network deployment module, cruise module, preprocessing module and monitoring module are used to collect and preprocess environmental data and biological data, use multimodal deep learning models to monitor ecological events, and dynamically adjust wireless Sensor network topology to improve data acquisition and processing efficiency.

Benefits of technology

It improves data collection and processing efficiency, enhances the protection capabilities of ecosystems, provides immediate management and decision-making support, and significantly promotes biodiversity conservation efforts.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119046881A_ABST
    Figure CN119046881A_ABST
Patent Text Reader

Abstract

The invention provides a biodiversity protection area monitoring system, and aims to improve ecological monitoring efficiency and protection capability. The system comprises the following modules: a network deployment module which deploys an intelligent sensor network in a target protection area and configures a plurality of wireless sensor nodes to collect environmental data, biological data and first time sequence data; and the cruising module is used for performing regular cruising by using the unmanned aerial vehicle and acquiring ground image data and second time sequence data. The preprocessing module is used for setting edge computing nodes at the edge of the network, preprocessing the obtained data and generating sensor and image preprocessing data; and the monitoring module is used for inputting the preprocessed data and the time sequence data into the trained ecological event monitoring model and outputting a monitoring result. By integrating multi-source data and an advanced processing technology, the efficiency of data acquisition and processing is improved, instant management and decision support are provided, and development of biodiversity protection work is remarkably promoted.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of ecosystem monitoring, and in particular to a biodiversity conservation area monitoring system. Background Art

[0002] Biodiversity conservation areas are important places for protecting natural resources and ecosystems. However, due to the limitation of personnel and technical resources, real-time monitoring and management of these areas has always been a challenge. Smart monitoring systems can help managers obtain ecological change information of protected areas in a timely manner and improve protection efficiency.

[0003] The prior art with application number 202311421382.0 discloses a biodiversity conservation area monitoring method, comprising the following steps: applying an integrated data fusion algorithm including a weighted average method and K-means clustering to fuse remote sensing data, geographic information system data, and ecological survey data to generate fused biodiversity data; based on the fused biodiversity data, using Mel frequency cepstrum coefficients to extract acoustic wave features, and using a support vector machine for pattern recognition to automatically identify and classify the sounds of different species in the conservation area to generate sound recognition and classification result data; using the sound recognition and classification result data in combination with genetic sample data, using haplogroup analysis and principal coordinate analysis to evaluate the genetic variability between populations, and monitoring the population status, genetics, and other characteristics of species in the conservation area; and Genetic diversity, genetic evolution direction, generate genetic data analysis report; based on the genetic data analysis report, apply spatial autocorrelation, graph theory network centrality indicators including degree centrality and betweenness centrality to construct and evaluate the spatial network model and habitat connectivity in the protected area, and generate a habitat connectivity assessment report; according to the habitat connectivity assessment report, build a big data analysis platform, apply data mining algorithms including random forest and K-means clustering, conduct biodiversity monitoring and analysis, and output biodiversity big data analysis results; use the biodiversity big data analysis results, combined with high-resolution remote sensing images, apply deep learning algorithms including convolutional neural networks and U-Net image segmentation methods, predict species distribution and habitat suitability in the protected area, and generate a species distribution prediction report. However, the ecosystem is dynamically changing, and the analysis of a single time point cannot accurately assess long-term trends and sudden changes, which has resulted in poor monitoring of current biodiversity protected areas. Summary of the invention

[0004] In order to overcome the deficiencies of the prior art, an object of the present invention is to provide a biodiversity conservation area monitoring system.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A biodiversity conservation area monitoring system comprising:

[0007] A network deployment module is used to deploy an intelligent sensor network in a target protection zone; the intelligent sensor network includes a plurality of wireless sensor nodes; each of the wireless sensor nodes is respectively arranged at a preset position of the target protection zone; the wireless sensor nodes are used to collect environmental data, biological data of the current location and first time series data of the wireless sensor nodes for data collection;

[0008] A patrol module, used to use a drone to regularly patrol the target protected area to obtain ground image data of the target protected area and second time series data of image acquisition by the drone;

[0009] A preprocessing module, used to set an edge computing node at the network edge in the smart sensor network, and use the edge computing node to preprocess the environmental data, biological data and ground image data respectively to obtain sensor preprocessing data and image preprocessing data;

[0010] A monitoring module is used to input the sensor preprocessing data, the first time series data, the image preprocessing data and the second time series data into a trained ecological event monitoring model to obtain monitoring results.

[0011] Preferably, the environmental data include: temperature, humidity, air quality index, light intensity, soil moisture, air pressure and precipitation; the biological data include: sound, movement and activity trajectory, population density and life sign indicators.

[0012] Preferably, it also includes:

[0013] The topology structure module is used to dynamically adjust the connection with other wireless sensor nodes according to the communication state of one of the wireless sensor nodes and the communication states of the surrounding wireless sensor nodes to form an adaptive network topology structure.

[0014] Preferably, the topology structure module includes:

[0015] A status sending unit, used to use each of the wireless sensor nodes to periodically send status information; the status information includes: signal strength, data transmission rate, packet loss rate, delay, energy consumption, moving speed and historical connection quality;

[0016] A state receiving unit, used to use each of the wireless sensor nodes to receive state information of surrounding nodes, and determine nodes to be connected and nodes to be disconnected among the surrounding nodes based on a machine learning algorithm;

[0017] The network construction unit is used to adjust the network connection of each of the wireless sensor nodes according to the nodes to be connected and the nodes to be disconnected to form the adaptive network topology structure.

[0018] Preferably, the preprocessing module comprises:

[0019] An image processing unit, used for preprocessing the ground image data by using an adaptive threshold filtering technique to obtain image preprocessing data;

[0020] The sensor processing unit is used to perform format conversion, data denoising and smoothing, and data filtering on the environmental data and the biological data in sequence to obtain the sensor preprocessing data.

[0021] Preferably, the image processing unit comprises:

[0022] A decomposition subunit, used for performing wavelet decomposition on the ground image data at multiple scales to obtain corresponding wavelet coefficients;

[0023] A threshold construction subunit, used to construct a filtering threshold according to the size and decomposition scale of the ground image data;

[0024] A model building subunit, used to build an image filtering model using the filtering threshold;

[0025] The filtering subunit is used to filter the ground image data using the image filtering model to obtain image preprocessing data.

[0026] Preferably, the calculation formula of the filtering threshold is:

[0027]

[0028] Among them, λ j represents the filtering threshold at the jth decomposition scale, σ j represents the standard deviation of the noise at the jth decomposition scale, N j represents the signal length at the jth decomposition scale, and j represents the decomposition scale.

[0029] Preferably, the calculation formula of the image filtering model is:

[0030]

[0031] Wherein, sign is the sign function, a is the first preset coefficient, b is the second preset coefficient, ω j,k represents the kth wavelet coefficient at the jth decomposition scale, Represents the wavelet coefficients after filtering.

[0032] Preferably, the method for constructing the ecological event monitoring model includes:

[0033] Collect multimodal data sets from sensor networks and drones; the multimodal data sets include historical environmental data, biological data, ground image data, and time series data corresponding to each data;

[0034] Annotating the multimodal dataset for known ecological events according to expert knowledge to obtain an annotated dataset;

[0035] Constructing a multimodal deep learning model; the multimodal deep learning model includes an input layer, a feature extraction network, a long short-term memory network, a fully connected network, a fusion layer and an output layer;

[0036] The labeled data set is input into the multimodal deep learning model for training to obtain the trained ecological event monitoring model.

[0037] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0038] The present invention provides a biodiversity conservation area monitoring system, comprising: a network deployment module, used to deploy an intelligent sensor network in a target conservation area; the intelligent sensor network comprises a plurality of wireless sensor nodes; each of the wireless sensor nodes is respectively arranged at respective preset positions of the target conservation area; the wireless sensor node is used to collect environmental data, biological data and first time series data of data collection by the wireless sensor node at the current location; a cruise module, used to use a drone to regularly cruise the target conservation area to obtain ground image data of the target conservation area and second time series data of image collection by the drone; a preprocessing module, used to set an edge computing node at the network edge of the intelligent sensor network, and use the edge computing node to preprocess the environmental data, biological data and ground image data respectively to obtain sensor preprocessing data and image preprocessing data; a monitoring module, used to input the sensor preprocessing data, the first time series data, the image preprocessing data and the second time series data into a trained ecological event monitoring model to obtain monitoring results. The present invention not only improves the efficiency of data collection and processing, but also enhances the protection capability of the ecosystem, provides instant management and decision support, and thus significantly promotes biodiversity conservation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0040] Figure 1 A schematic diagram of the system structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] The purpose of the present invention is to provide a biodiversity conservation area monitoring system, which not only improves the efficiency of data collection and processing, but also enhances the ability to protect ecosystems, provides immediate management and decision-making support, and thus significantly promotes biodiversity conservation.

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] Figure 1 A schematic diagram of the system structure provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a biodiversity conservation area monitoring system, comprising:

[0045] A network deployment module is used to deploy an intelligent sensor network in a target protection zone; the intelligent sensor network includes a plurality of wireless sensor nodes; each of the wireless sensor nodes is respectively arranged at a preset position of the target protection zone; the wireless sensor nodes are used to collect environmental data, biological data of the current location and first time series data of the wireless sensor nodes for data collection;

[0046] A patrol module, used to use a drone to regularly patrol the target protected area to obtain ground image data of the target protected area and second time series data of image acquisition by the drone;

[0047] A preprocessing module, used to set an edge computing node at the network edge in the smart sensor network, and use the edge computing node to preprocess the environmental data, biological data and ground image data respectively to obtain sensor preprocessing data and image preprocessing data;

[0048] A monitoring module is used to input the sensor preprocessing data, the first time series data, the image preprocessing data and the second time series data into a trained ecological event monitoring model to obtain monitoring results.

[0049] Preferably, the environmental data include: temperature, humidity, air quality index, light intensity, soil moisture, air pressure and precipitation; the biological data include: sound, movement and activity trajectory, population density and life sign indicators.

[0050] Specifically, temperature refers to measuring the current air temperature or water temperature to monitor climate change and its impact on species habitats. Humidity refers to monitoring the humidity level in the air to observe its impact on plant and animal habitat conditions. Air quality indicators include gas concentrations (such as CO2, NO x 、SO x ), monitoring the impact of human activities and natural events on air quality. Light intensity specifically refers to measuring the intensity of light, which helps understand plant growth conditions, sunshine hours, etc. Soil moisture specifically determines the moisture content of the soil. This indicator of soil moisture is critical for plant growth and health monitoring. Air pressure specifically monitors changes in air pressure to predict weather patterns or climate change. Precipitation specifically collects precipitation data to help understand the impact on water resources in ecosystems.

[0051] Furthermore, sound specifically captures animal sounds and analyzes biodiversity and species activities. It can be used for automatic identification and classification of biological species. Movement and activity specifically refers to detecting the activity level of animals or other organisms through accelerometers or other motion sensors. Population density is the number of animals obtained by sensors to estimate the density and number changes of different species in a specific area. Vital sign indicators include heart rate and body temperature (where possible) to monitor the health status of individual animals. Acoustic habitat characteristics specifically capture the background acoustic environment for analysis and differentiation of different acoustic environment characteristics.

[0052] By integrating these environmental and biological data and combining them with time series information, this embodiment can analyze ecological changes in real time and identify potential ecological events, such as habitat destruction, abnormal species migration patterns, and environmental pollution. With edge computing and ecological event monitoring models, data processing and analysis can be completed near the sensor network, thereby improving the system's response speed and accuracy.

[0053] Preferably, it also includes:

[0054] The topology structure module is used to dynamically adjust the connection with other wireless sensor nodes according to the communication state of one of the wireless sensor nodes and the communication states of the surrounding wireless sensor nodes to form an adaptive network topology structure.

[0055] Specifically, in this embodiment, by dynamically adjusting the connection according to the communication status of the node, the system can quickly respond to environmental changes (such as node failure, movement, or signal quality degradation), thereby improving the adaptability and robustness of the network.

[0056] Preferably, the topology structure module includes:

[0057] A status sending unit, used to use each of the wireless sensor nodes to periodically send status information; the status information includes: signal strength, data transmission rate, packet loss rate, delay, energy consumption, moving speed and historical connection quality;

[0058] A state receiving unit, used to use each of the wireless sensor nodes to receive state information of surrounding nodes, and determine nodes to be connected and nodes to be disconnected among the surrounding nodes based on a machine learning algorithm;

[0059] The network construction unit is used to adjust the network connection of each of the wireless sensor nodes according to the nodes to be connected and the nodes to be disconnected to form the adaptive network topology structure.

[0060] Optionally, this embodiment periodically sends and receives status information so that the system can grasp the network status of each node in real time. This feedback mechanism of this embodiment helps to reduce communication delays, increase data transmission rates, and reduce energy consumption. In addition, this embodiment has an intelligent decision-making mechanism, which can more intelligently select nodes to be connected and disconnected by analyzing the status information of surrounding nodes through machine learning algorithms. The intelligent decision-making of this embodiment can improve the connection quality and the overall performance of the network.

[0061] Furthermore, the embodiment can efficiently select connection nodes to reduce energy consumption and extend the service life of sensor nodes, especially in the case of limited resources.

[0062] Furthermore, by optimizing the network topology, this embodiment can significantly improve the success rate of data transmission, reduce packet loss rate and delay, and ensure the real-time and accuracy of data transmission. In addition, this embodiment enables the wireless sensor network to flexibly adapt to different environments and conditions, meeting the needs of different application scenarios.

[0063] In addition, as new nodes are added, this embodiment can dynamically adjust to maintain network efficiency and adapt to expansion needs without causing performance degradation. When a node fails, dynamic adjustment of links can quickly find alternative paths to maintain network stability and integrity.

[0064] In summary, this embodiment significantly improves the overall performance of the wireless sensor network through a dynamic adaptive method, improves the efficiency and reliability of data transmission, and promotes the effectiveness of the biodiversity monitoring system, especially in complex and changing environments such as target protected areas.

[0065] Preferably, the preprocessing module comprises:

[0066] An image processing unit, used for preprocessing the ground image data by using an adaptive threshold filtering technique to obtain image preprocessing data;

[0067] The sensor processing unit is used to perform format conversion, data denoising and smoothing, and data filtering on the environmental data and the biological data in sequence to obtain the sensor preprocessing data.

[0068] Preferably, the image processing unit comprises:

[0069] A decomposition subunit, used for performing wavelet decomposition on the ground image data at multiple scales to obtain corresponding wavelet coefficients;

[0070] A threshold construction subunit, used to construct a filtering threshold according to the size and decomposition scale of the ground image data;

[0071] A model building subunit, used to build an image filtering model using the filtering threshold;

[0072] The filtering subunit is used to filter the ground image data using the image filtering model to obtain image preprocessing data.

[0073] Preferably, the calculation formula of the filtering threshold is:

[0074]

[0075] Among them, λ j represents the filtering threshold at the jth decomposition scale, σ j represents the standard deviation of the noise at the jth decomposition scale, N j represents the signal length at the jth decomposition scale, and j represents the decomposition scale.

[0076] Specifically, when using wavelet denoising, the value of the wavelet threshold has an important influence on the final denoising effect. The wavelet threshold should be able to separate the wavelet coefficients of the noise and the original image signal to the greatest extent. As the number of wavelet decomposition layers increases, the wavelet coefficient area where the noise is located should gradually decrease. Most of the wavelet coefficients decomposed at a higher level come from the original image signal. Therefore, the present invention constructs a self-transformed filtering threshold based on the decomposition scale.

[0077] Furthermore, it can be seen from the above filter threshold formula that as the decomposition scale increases, the threshold of each layer decreases accordingly, which is in line with the law of wavelet denoising, and thus can improve the performance of wavelet denoising.

[0078] Preferably, the calculation formula of the image filtering model is:

[0079]

[0080] Wherein, sign is the sign function, a is the first preset coefficient, b is the second preset coefficient, ω j,k represents the kth wavelet coefficient at the jth decomposition scale, Represents the wavelet coefficients after filtering.

[0081] Preferably, the method for constructing the ecological event monitoring model includes:

[0082] Collect multimodal data sets from sensor networks and drones; the multimodal data sets include historical environmental data, biological data, ground image data, and time series data corresponding to each data;

[0083] Annotating the multimodal dataset for known ecological events according to expert knowledge to obtain an annotated dataset;

[0084] Constructing a multimodal deep learning model; the multimodal deep learning model includes an input layer, a feature extraction network, a long short-term memory network, a fully connected network, a fusion layer and an output layer;

[0085] The labeled data set is input into the multimodal deep learning model for training to obtain the trained ecological event monitoring model.

[0086] Specifically, this embodiment first performs data collection and preparation, and the specific steps are as follows:

[0087] 1. Data collection: Collect multimodal data from sensor networks and drones, including historical environmental data, biological data, corresponding time series data, and ground image data.

[0088] 2. Data annotation: Annotate data sets for known ecological events based on expert knowledge. This includes identifying known ecological events (such as species emergence, habitat changes, etc.).

[0089] 3. Data cleaning and enhancement: Clean up noisy data and improve data quality. Use techniques such as data interpolation and synthesizing minority class samples to enhance the diversity of the data set.

[0090] Furthermore, this embodiment then establishes a model structure, and the specific steps are as follows:

[0091] 1. Multimodal Deep Learning Architecture

[0092] Input layer: Receives different types of data channels – numerical data from sensors and image data.

[0093] Feature extraction network, specifically including: convolutional neural network, recurrent neural network (RNN) or long short-term memory network, and fully connected network; convolutional neural network (CNN) is used to process and extract spatial features in ground image data; recurrent neural network (RNN) or long short-term memory network (LSTM) is used to process time series data and capture temporal dynamic features; fully connected network (DNN) is used to extract complex nonlinear features from numerical data (such as environmental and biological sensor data).

[0094] Fusion layer: The features of the above channels are integrated and feature fusion is achieved through attention mechanism or multimodal fusion network.

[0095] Output layer: classification or regression output, used to predict and identify ecological events.

[0096] 2. Model training process

[0097] Construct a loss function and select an appropriate loss function (such as cross entropy loss for classification problems) to analyze model performance.

[0098] Determine the optimization algorithm and use optimization algorithms such as Adam or SGD to iteratively optimize the model parameters through gradient descent.

[0099] Split the data into training, validation, and test sets to evaluate the generalization ability of the model.

[0100] Hyperparameter tuning: Optimize the model's hyperparameters through methods such as cross-validation or Bayesian optimization.

[0101] Optionally, the monitoring results of this embodiment include outputting the probability or type of ecological events in a specified area, including species migration, the impact of abnormal weather, habitat changes, etc., and based on the input time series data, the model can predict the future changing trends of certain ecological events or characteristics.

[0102] Furthermore, this embodiment also provides an intuitive data visualization interface to display monitoring results, such as event heat maps and time dynamic change curves. In addition, this embodiment integrates an automatic early warning subsystem. When a major ecological anomaly is detected, the system can immediately send an alarm message to relevant management personnel to facilitate rapid response.

[0103] The beneficial effects of the present invention are as follows:

[0104] By combining different types of data and machine learning techniques, the ecological event monitoring model of the present invention can efficiently analyze and detect ecological changes in protected areas, providing reliable technical support for biodiversity conservation.

[0105] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0106] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A biodiversity conservation area monitoring system, characterized in that: include: A network deployment module is used to deploy an intelligent sensor network in a target protection zone; the intelligent sensor network includes a plurality of wireless sensor nodes; each of the wireless sensor nodes is respectively arranged at a preset position of the target protection zone; the wireless sensor nodes are used to collect environmental data, biological data of the current location and first time series data of the wireless sensor nodes for data collection; A patrol module, used to use a drone to regularly patrol the target protected area to obtain ground image data of the target protected area and second time series data of image acquisition by the drone; A preprocessing module, used to set an edge computing node at the network edge in the smart sensor network, and use the edge computing node to preprocess the environmental data, biological data and ground image data respectively to obtain sensor preprocessing data and image preprocessing data; A monitoring module is used to input the sensor preprocessing data, the first time series data, the image preprocessing data and the second time series data into a trained ecological event monitoring model to obtain monitoring results.

2. The biodiversity conservation area monitoring system according to claim 1, characterized in that: The environmental data include: temperature, humidity, air quality index, light intensity, soil moisture, air pressure and precipitation; the biological data include: sound, movement and activity trajectory, population density and life sign index.

3. The biodiversity conservation area monitoring system according to claim 1, characterized in that: Also includes: The topology structure module is used to dynamically adjust the connection with other wireless sensor nodes according to the communication state of one of the wireless sensor nodes and the communication states of the surrounding wireless sensor nodes to form an adaptive network topology structure.

4. The biodiversity conservation area monitoring system according to claim 3, characterized in that: The topology structure module includes: A status sending unit, used to use each of the wireless sensor nodes to periodically send status information; the status information includes: signal strength, data transmission rate, packet loss rate, delay, energy consumption, moving speed and historical connection quality; A state receiving unit, used to use each of the wireless sensor nodes to receive state information of surrounding nodes, and determine nodes to be connected and nodes to be disconnected among the surrounding nodes based on a machine learning algorithm; The network construction unit is used to adjust the network connection of each of the wireless sensor nodes according to the nodes to be connected and the nodes to be disconnected to form the adaptive network topology structure.

5. The biodiversity conservation area monitoring system according to claim 1, characterized in that: The preprocessing module comprises: An image processing unit, used for preprocessing the ground image data by using an adaptive threshold filtering technique to obtain image preprocessing data; The sensor processing unit is used to perform format conversion, data denoising and smoothing, and data filtering on the environmental data and the biological data in sequence to obtain the sensor preprocessing data.

6. The biodiversity conservation area monitoring system according to claim 5, characterized in that: The image processing unit comprises: A decomposition subunit, used for performing wavelet decomposition on the ground image data at multiple scales to obtain corresponding wavelet coefficients; A threshold construction subunit, used to construct a filtering threshold according to the size and decomposition scale of the ground image data; A model building subunit, used to build an image filtering model using the filtering threshold; The filtering subunit is used to filter the ground image data using the image filtering model to obtain image preprocessing data.

7. The biodiversity conservation area monitoring system according to claim 6, characterized in that: The calculation formula of the filtering threshold is: Among them, λ j represents the filtering threshold at the jth decomposition scale, σ j represents the standard deviation of the noise at the jth decomposition scale, N j represents the signal length at the jth decomposition scale, and j represents the decomposition scale.

8. The biodiversity conservation area monitoring system according to claim 7, characterized in that: The calculation formula of the image filtering model is: Wherein, sign is the sign function, a is the first preset coefficient, b is the second preset coefficient, ω j,k represents the kth wavelet coefficient at the jth decomposition scale, Represents the wavelet coefficients after filtering.

9. The biodiversity conservation area monitoring system according to claim 8, characterized in that: The method for constructing the ecological event monitoring model includes: Collect multimodal data sets from sensor networks and drones; the multimodal data sets include historical environmental data, biological data, ground image data, and time series data corresponding to each data; Annotating the multimodal dataset for known ecological events according to expert knowledge to obtain an annotated dataset; Constructing a multimodal deep learning model; the multimodal deep learning model includes an input layer, a feature extraction network, a long short-term memory network, a fully connected network, a fusion layer and an output layer; The labeled data set is input into the multimodal deep learning model for training to obtain the trained ecological event monitoring model.

Citation Information

Patent Citations

  • Biodiversity protection area monitoring method and system

    CN117171602A

  • Intelligent monitoring method and system for watershed ecosystem

    CN118381638A

  • Basin ecosystem intelligent monitoring method and system based on remote sensing technology

    CN118429826A

  • Unattended water environment and water ecology monitoring system and method

    CN118443899A

  • Edge server, edge computing system and communication method thereof

    CN118509949A

Cited By

  • Biodiversity monitoring system

    CN120489211A

  • Automatic waterfowl biodiversity monitoring and collecting device

    CN120760790A