Marine ecological evaluation and analysis system based on big data

Abstract

The present invention relates to the field of data processing technology, and specifically discloses a marine ecological evaluation and analysis system based on big data, including a data acquisition module, a data processing module, a data analysis module and a comprehensive evaluation module. The present invention can predict and analyze the distribution area of ​​marine biological species, can identify potential invasive species in marine biological species according to the predicted species distribution map analysis, can analyze the marine water pollution rate, the marine landform change rate and the marine resource change rate, can identify and analyze sea surface image data, analyze marine remote sensing data, and can evaluate the marine ecological risk level after comprehensively analyzing the risk assessment index of marine ecology based on various data; the present invention can more comprehensively understand the status of the marine ecosystem by analyzing a variety of data, thereby more accurately evaluating the risk level of the marine ecology.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and specifically discloses a marine ecological evaluation and analysis system based on big data. Background Art

[0002] With the continuous advancement of information technology, big data technology is becoming increasingly mature, providing a scientific basis for decision-making. In the field of marine ecological evaluation and analysis, big data technology also has great application potential. The ocean is one of the most important ecosystems on the earth. Marine ecology includes the relationship between marine organisms and between marine organisms and their marine environment. The ocean provides a large amount of resources for humans every year. It is increasingly important to use big data technology, machine learning and other technical means to analyze marine ecological data and evaluate marine ecology.

[0003] The existing technology has the following problems: first, it is difficult to predict and analyze the distribution areas of marine biological species; second, it is difficult to identify potential invasive species among marine biological species based on the predicted species distribution map analysis; then, it is difficult to analyze the marine water pollution rate, the marine landform change rate and the marine resource change rate; then, it is difficult to identify and analyze sea surface image data and difficult to analyze marine remote sensing data; finally, it is difficult to evaluate the marine ecological risk level after comprehensively analyzing the marine ecological risk assessment index based on various data. Summary of the invention

[0004] In view of the above deficiencies in the prior art, the present invention provides a marine ecological evaluation and analysis system based on big data, which can predict and analyze the distribution areas of marine biological species, can identify potential invasive species in marine biological species according to the predicted species distribution map analysis, can analyze the marine water pollution rate, the marine landform change rate and the marine resource change rate, can identify and analyze sea surface image data, analyze marine remote sensing data, and can evaluate the marine ecological risk level after comprehensively analyzing the marine ecological risk assessment index based on various data.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] The marine ecological evaluation and analysis system based on big data includes the following modules:

[0007] Data acquisition module: collects marine ecological data, sea surface image data and marine remote sensing data;

[0008] Data processing module: pre-processing the collected marine ecological data, including cleaning and standardization; pre-processing the collected sea surface image data, including denoising and image enhancement; pre-processing the collected marine remote sensing data, including denoising, geometric correction and radiation correction;

[0009] Data analysis module: predict and analyze the distribution area of ​​marine biological species based on the collected and processed marine biological data and marine environmental data, and calculate the marine biological niche overlap index based on the prediction results; identify potential invasive species among marine biological species based on the predicted species distribution map, and calculate the impact rate of potential invasive species based on the analysis results; calculate the marine water quality pollution rate, marine landform change rate and marine resource change rate based on the marine ecological data analysis; identify and determine whether red tide areas and green tide areas appear based on the collected and processed sea surface image data, and calculate the sea surface pollution rate based on the identification results; extract features based on the collected and processed marine remote sensing data, and calculate the preliminary productivity coefficient and coral reef bleaching change rate based on the extracted feature parameters;

[0010] Comprehensive evaluation module: The risk assessment index of marine ecology is calculated based on a comprehensive analysis of the marine biological niche overlap index, the impact rate of potential invasive species, the marine water quality pollution rate, the marine landform change rate, the marine resource change rate, the sea surface pollution rate, the preliminary productivity coefficient and the coral reef bleaching change rate, and the marine ecology is evaluated based on the analysis results.

[0011] Furthermore, marine ecological data are collected, including: marine biological data, marine water quality data, marine resource data and marine environmental data; the marine biological data include: the number and type of marine biological species; the marine water quality data include: seawater salinity and pollutant concentration; the marine resource data include: fishery resources, mineral resources and renewable energy resources; the marine environmental data include: temperature, humidity, wind speed, dissolved oxygen and sediment quality.

[0012] Further, the distribution area of ​​marine biological species is predicted and analyzed based on the collected and processed marine biological data and marine environmental data, and the marine biological niche overlap index is calculated based on the prediction results, including the following steps:

[0013] Based on the collected and processed marine biological data and marine environmental data, the maximum entropy model is used to predict and analyze the distribution areas of marine biological species;

[0014] Input the marine biological data and marine environmental data into the maximum entropy model for model training; according to the trained maximum entropy model, the marine ecological evaluation and analysis system based on big data collects and processes the marine biological data and marine environmental data in real time, inputs the processed marine biological data and marine environmental data into the trained maximum entropy model, predicts the distribution area of ​​marine biological species, and outputs the prediction results; according to the prediction results, converts the probability value of species occurrence in each grid unit into a species distribution map;

[0015] According to the prediction results, by analyzing the formula:

[0016] The marine biological niche overlap index F is obtained; where H represents the total number of marine resources, P ih represents the frequency of marine species i under marine resource h, P jh represents the frequency of marine species j under marine resource h.

[0017] Further, the potential invasive species in the marine biological species are analyzed and identified according to the predicted species distribution map, and the impact rate of the potential invasive species is calculated according to the analysis results, including the following steps:

[0018] Based on the predicted species distribution map, the predicted species distribution map is compared with the historical species distribution map, and species that do not appear in the historical species distribution map but appear in the predicted species distribution map are identified and marked as potential invasive species;

[0019] By analyzing the formula:

[0020] The impact rate E of potential invasive species is obtained; where S represents the total number of marine species and f represents the total number of potential invasive species.

[0021] Further, the marine water pollution rate, marine landform change rate and marine resource change rate are calculated based on the marine ecological data analysis, including the following steps:

[0022] By analyzing the formula:

[0023] Get the ocean water pollution rate R; where C i represents the concentration of the i-th pollutant in the marine water quality data, Z i represents the standard limit of the concentration of the i-th pollutant, and m represents the total number of pollutants detected;

[0024] The historical and real-time changes in sea level and marine resources at time point t;

[0025] By analyzing the formula:

[0026] The ocean landform change rate M is obtained; where ΔH represents the change in sea level height, and Δt represents the length of the time period;

[0027] By analyzing the formula:

[0028] The ocean resource change rate Y is obtained; wherein ΔQ represents the change in ocean resources, and Δt represents the length of the time period.

[0029] Further, according to the collected and processed sea surface image data, whether a red tide area and a green tide area appear is identified and judged, and the sea surface pollution rate is calculated according to the identification result analysis, including the following steps:

[0030] Based on the collected and processed sea surface image data, the convolutional neural network model in the deep learning model is used to identify and determine whether red tide areas and green tide areas appear in the sea surface image data;

[0031] Label the red tide areas and green tide areas in the sea surface image data; input the labeled image data into the convolutional neural network model for training;

[0032] According to the trained convolutional neural network model, the sea surface image data collected and processed in real time by the marine ecological evaluation and analysis system based on big data is input into the trained convolutional neural network model for automatic recognition, and it is determined whether the sea surface image data collected in real time contains red tide areas and green tide areas, and the recognition results are output;

[0033] When it is determined that red tide areas and green tide areas appear in the sea surface image data collected in real time, the coverage areas of the red tide areas and green tide areas in the sea surface image data are extracted; otherwise, the sea surface image data is collected again in real time for identification and judgment;

[0034] By analyzing the formula:

[0035] The sea surface pollution rate G is obtained; where Red represents the coverage area of ​​the red tide area, Green represents the coverage area of ​​the green tide area, Area represents the total area of ​​the sea surface, and ε Red represents the impact factor of the red tide area, ε Green Represents the influencing factor of the green tide area.

[0036] Further, feature extraction is performed based on the collected and processed ocean remote sensing data, and preliminary productivity coefficients and coral reef bleaching change rates are calculated based on the extracted feature parameters, including the following steps:

[0037] Extract characteristic parameters from the collected and processed ocean remote sensing data by image segmentation and classification methods, wherein the characteristic parameters include: ocean surface temperature and chlorophyll concentration; the ocean remote sensing data includes: multispectral image data and hyperspectral image data;

[0038] Based on the collected and processed ocean remote sensing data, deep learning models are used to analyze and identify phytoplankton areas and coral reef bleaching areas in the ocean remote sensing data;

[0039] Label the hyperspectral image data of phytoplankton areas and coral reef bleaching areas, and input the labeled data into the deep learning model for training;

[0040] According to the trained deep learning model, the marine remote sensing data collected and processed in real time by the marine ecological evaluation and analysis system based on big data is input into the trained deep learning model to automatically identify the phytoplankton areas and coral reef bleaching areas and output the identification results;

[0041] According to the recognition result, the ocean remote sensing data collected historically at time point t is represented as a first remote sensing image, and the ocean remote sensing data collected in real time at time point t is represented as a second remote sensing image;

[0042] By analyzing the formula:

[0043] The preliminary productivity coefficient β is obtained; wherein A represents the area of ​​phytoplankton in the first remote sensing image, B represents the area of ​​phytoplankton in the second remote sensing image, N represents the chlorophyll concentration, η represents the conversion coefficient, η=3.7, and T represents the sea surface temperature;

[0044] By analyzing the formula:

[0045] Get the coral reef bleaching change rate C b ; Wherein, C1 represents the area of ​​coral reef bleaching in the first remote sensing image, and C2 represents the area of ​​coral reef bleaching in the second remote sensing image.

[0046] Furthermore, the comprehensive evaluation module comprises the following steps:

[0047] By analyzing the formula:

[0048]

[0049] The risk assessment index φ is obtained; among them, F represents the marine biological niche overlap index, E represents the impact rate of potential invasive species, R represents the marine water pollution rate, M represents the marine landform change rate, Y represents the marine resource change rate, G represents the sea surface pollution rate, β represents the initial productivity coefficient, C b represents the change rate of coral reef bleaching; W1, W2, W3, W4, W5, W6, W7 and W8 represent the weight coefficients of marine biological niche overlap index, potential invasive species impact rate, marine water quality pollution rate, marine geomorphological change rate, marine resource change rate, sea surface pollution rate, initial productivity coefficient and coral reef bleaching change rate, respectively;

[0050] Set thresholds and classify them according to the results of the risk assessment index to evaluate the risk level of marine ecology;

[0051] When φ<25%, the risk level of marine ecology is low risk;

[0052] When 25%≤φ≤60%, the risk level of marine ecology is medium risk;

[0053] When φ>60%, the risk level of marine ecology is high risk.

[0054] The beneficial effects of the present invention are:

[0055] 1. The present invention predicts and analyzes the distribution area of ​​marine biological species through the maximum entropy model, comprehensively considers the ecological niche of species, environmental factors and interactions between species, so as to accurately predict the distribution area of ​​species; based on the predicted species distribution map, the present invention uses advanced image processing and data analysis technology to identify potential invasive species in marine biological species;

[0056] 2. The present invention can accurately identify red tide and green tide areas in sea surface images by training a deep learning model, thereby improving the accuracy of red tide and green tide identification; by extracting features from marine remote sensing data and analyzing the extracted feature parameters, the preliminary productivity coefficient and coral reef bleaching change rate are obtained; the deep learning technology is used to improve the efficiency of monitoring and reduce the time and cost of manual interpretation;

[0057] 3. The present invention can analyze a variety of data including satellite remote sensing data, marine ecological data, and ocean surface image data to more comprehensively understand the status of the marine ecosystem and more accurately assess its risks. It can also use big data technology to achieve real-time collection, processing, and analysis of marine ecological data, thereby improving the timeliness of risk assessment. It can also comprehensively analyze the risk assessment index of marine ecology through various data to evaluate the analysis level of marine ecology, thereby achieving a comprehensive evaluation of marine ecology and improving accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0059] The present invention will be described and illustrated in detail below in conjunction with the accompanying drawings.

[0060] Example 1

[0061] like Figure 1 As shown, the marine ecological evaluation and analysis system based on big data includes the following modules:

[0062] Data acquisition module: collects marine ecological data, sea surface image data and marine remote sensing data;

[0063] Data processing module: pre-processing the collected marine ecological data, including cleaning and standardization; pre-processing the collected sea surface image data, including denoising and image enhancement; pre-processing the collected marine remote sensing data, including denoising, geometric correction and radiation correction;

[0064] Data analysis module: predict and analyze the distribution area of ​​marine biological species based on the collected and processed marine biological data and marine environmental data, and calculate the marine biological niche overlap index based on the prediction results; identify potential invasive species among marine biological species based on the predicted species distribution map, and calculate the impact rate of potential invasive species based on the analysis results; calculate the marine water quality pollution rate, marine landform change rate and marine resource change rate based on the marine ecological data analysis; identify and determine whether red tide areas and green tide areas appear based on the collected and processed sea surface image data, and calculate the sea surface pollution rate based on the identification results; extract features based on the collected and processed marine remote sensing data, and calculate the preliminary productivity coefficient and coral reef bleaching change rate based on the extracted feature parameters;

[0065] Comprehensive evaluation module: The risk assessment index of marine ecology is calculated based on a comprehensive analysis of the marine biological niche overlap index, the impact rate of potential invasive species, the marine water quality pollution rate, the marine landform change rate, the marine resource change rate, the sea surface pollution rate, the preliminary productivity coefficient and the coral reef bleaching change rate, and the marine ecology is evaluated based on the analysis results.

[0066] Specifically, marine sensors including temperature, salinity, dissolved oxygen and other sensors and marine observatory equipment are used to collect marine ecological data in real time. Sea surface image data is collected through satellites, drones or ground monitoring equipment. Satellite remote sensing technology is used to obtain multi-spectral, hyperspectral and other image information of the ocean surface for analyzing the macroscopic characteristics of marine ecology. Marine ecological data preprocessing includes: removing duplicate, invalid or erroneous data; converting data into a unified format and unit for subsequent analysis. Sea surface image data preprocessing includes: using filtering algorithms to remove noise and clutter in images; improving image clarity and readability through contrast enhancement, sharpening and other processing. Marine remote sensing data preprocessing includes: using filtering algorithms to remove noise in remote sensing images; through geometric transformation, the geometric shape and position of remote sensing images are consistent with geographic coordinates; converting the brightness value of remote sensing images into radiation values ​​to accurately reflect the radiation characteristics of the ground objects. Based on the collected marine biological and environmental data, statistical analysis and machine learning algorithms are used to predict the distribution area of ​​marine biological species. Analyze and calculate the niche overlap index of marine organisms to evaluate the ecological competition relationship between different species. Identify potential invasive species and calculate their impact rate. Calculate the marine water pollution rate and evaluate the water quality based on marine ecological data. Analyze the marine landform change rate and monitor changes in coastlines, seabed topography, etc. Calculate the marine resource change rate and evaluate the development and utilization of fisheries, minerals and other resources. Use image processing algorithms to identify and determine red tide and green tide areas. Analyze and calculate the sea surface pollution rate and evaluate the pollution status of the sea surface environment. After extracting characteristic parameters from remote sensing images, analyze and calculate the preliminary productivity coefficient to evaluate the productivity level of the marine ecosystem; monitor the change rate of coral reef bleaching and evaluate the change rate of coral reef bleaching. Calculate the risk assessment index of marine ecology using a comprehensive evaluation method based on various data, and grade and evaluate the marine ecological status according to the size of the risk assessment index.

[0067] Collect marine ecological data, including: marine biological data, marine water quality data, marine resource data and marine environmental data; the marine biological data include: the number and type of marine biological species; the marine water quality data include: seawater salinity and pollutant concentration; the marine resource data include: fishery resources, mineral resources and renewable energy resources; the marine environmental data include: temperature, humidity, wind speed, dissolved oxygen and sediment quality.

[0068] Specifically, use divers, underwater robots and other equipment to observe and count marine life, and collect marine life data including: species number and species type; set up water quality monitoring stations or sensors in the target sea area to monitor seawater salinity, pollutant concentration and other indicators in real time. Record the seawater salinity and pollutant concentration in the marine water quality data according to the set collection frequency. Obtain the distribution and reserve information of fishery resources and mineral resources through methods such as fishery surveys and mineral resource exploration, and use renewable energy assessment models to estimate the potential of renewable energy resources in the sea area. Set up meteorological stations or environmental sensors in the target sea area to monitor temperature, humidity, wind speed and other indicators in real time. Use equipment such as dissolved oxygen meters and sediment samplers to collect dissolved oxygen and sediment quality data.

[0069] The distribution areas of marine biological species are predicted and analyzed based on the collected and processed marine biological data and marine environmental data, and the marine biological niche overlap index is calculated based on the prediction results, including the following steps:

[0070] Based on the collected and processed marine biological data and marine environmental data, the maximum entropy model is used to predict and analyze the distribution areas of marine biological species;

[0071] Input the marine biological data and marine environmental data into the maximum entropy model for model training; according to the trained maximum entropy model, the marine ecological evaluation and analysis system based on big data collects and processes the marine biological data and marine environmental data in real time, inputs the processed marine biological data and marine environmental data into the trained maximum entropy model, predicts the distribution area of ​​marine biological species, and outputs the prediction results; according to the prediction results, converts the probability value of species occurrence in each grid unit into a species distribution map;

[0072] According to the prediction results, by analyzing the formula:

[0073] The marine biological niche overlap index F is obtained; where H represents the total number of marine resources, P ih represents the frequency of marine species i under marine resource h, P jh represents the frequency of marine species j under marine resource h.

[0074] Specifically, the maximum entropy model is a machine learning algorithm based on information entropy, which is suitable for dealing with species distribution prediction problems. The cleaned and standardized marine biological data and marine environmental data are input into the maximum entropy model. The maximum entropy model is trained using the training data set, and the model parameters are adjusted to minimize the prediction error. The marine ecological evaluation and analysis system based on big data collects marine biological data and marine environmental data in real time and cleans and standardizes the real-time data. The processed real-time data is input into the trained maximum entropy model to predict the distribution area of ​​marine biological species. The prediction results are output, including the probability value of the species in each grid unit. According to the prediction results, the probability value of the species in each grid unit is converted into a species distribution map. Different colors or grayscales can be used to represent different probability values ​​to form an intuitive species distribution map. The niche overlap index is calculated by a formula, which is used to measure the overlap degree of different marine biological species on marine resources. According to the species distribution map and marine resource data, the frequency of each species under each marine resource is calculated. Substitute the calculated frequency into the niche overlap index formula to calculate the niche overlap index between different marine species; analyze the size and changing trend of the niche overlap index to understand the degree of overlap and competitive relationship between different marine species in marine resources.

[0075] Identify potential invasive species among marine biological species based on predicted species distribution map analysis, and calculate the impact rate of potential invasive species based on the analysis results, including the following steps:

[0076] Based on the predicted species distribution map, the predicted species distribution map is compared with the historical species distribution map, and species that do not appear in the historical species distribution map but appear in the predicted species distribution map are identified and marked as potential invasive species;

[0077] By analyzing the formula:

[0078] The impact rate E of potential invasive species is obtained; where S represents the total number of marine species and f represents the total number of potential invasive species.

[0079] Specifically, obtain a predicted species distribution map, which is based on the maximum entropy model or other prediction models and shows the possible distribution of marine species in the future. Obtain a historical species distribution map, which is based on past observation data or records and shows the actual distribution of marine species in the past. Overlay or compare the predicted species distribution map with the historical species distribution map to observe the differences between the two. Pay special attention to those species that do not appear in the historical species distribution map but appear in the predicted species distribution map. Mark the newly appeared species in the above difference species that meet the characteristics of invasive species as potential invasive species. The characteristics of invasive species may include rapid reproduction, wide distribution, and damage to the ecological environment. The impact rate formula is used to quantify the impact of potential invasive species on marine biodiversity and ecosystems. According to the predicted species distribution map or historical species distribution map, count the total number of marine species. According to the identified potential invasive species, count their total number. Substitute the above data into the formula to calculate the impact rate of potential invasive species. According to the calculated impact rate, analyze the degree of threat that potential invasive species may cause to marine biodiversity and ecosystems.

[0080] The calculation of the marine water pollution rate, marine landform change rate and marine resource change rate based on the marine ecological data analysis includes the following steps:

[0081] By analyzing the formula:

[0082] Get the ocean water pollution rate R; where C i represents the concentration of the i-th pollutant in the marine water quality data, Z i represents the standard limit of the concentration of the i-th pollutant, and m represents the total number of pollutants detected;

[0083] The historical and real-time changes in sea level and marine resources at time point t;

[0084] By analyzing the formula:

[0085] The ocean landform change rate M is obtained; where ΔH represents the change in sea level height, and Δt represents the length of the time period;

[0086] By analyzing the formula:

[0087] The ocean resource change rate Y is obtained; wherein ΔQ represents the change in ocean resources, and Δt represents the length of the time period.

[0088] Specifically, collect marine water quality data, including the concentration of various pollutants, including chemical oxygen demand, inorganic nitrogen, inorganic phosphorus and dissolved oxygen. Obtain marine water quality standard limits, which are usually set by environmental protection agencies or international organizations to assess whether water quality meets the standards, and match various pollutant concentration data with the corresponding water quality standard limits. Through comprehensive analysis and calculation, obtain pollution rate analysis to determine whether the water pollution is serious and whether there is a potential pollution risk. The historical and real-time sea level height changes and marine resource changes collected at the same time point t, analyze the size and trend of the landform change rate within the time period, and determine whether the marine landform is stable. According to the marine resource data, analyze the size and trend of the resource change rate within the time period to determine whether the development and utilization of marine resources is sustainable and whether there is a potential risk of resource depletion.

[0089] According to the collected and processed sea surface image data, whether a red tide area and a green tide area appear is identified and judged, and the sea surface pollution rate is calculated according to the identification result analysis, including the following steps:

[0090] Based on the collected and processed sea surface image data, the convolutional neural network model in the deep learning model is used to identify and determine whether red tide areas and green tide areas appear in the sea surface image data;

[0091] Label the red tide areas and green tide areas in the sea surface image data; input the labeled image data into the convolutional neural network model for training;

[0092] According to the trained convolutional neural network model, the sea surface image data collected and processed in real time by the marine ecological evaluation and analysis system based on big data is input into the trained convolutional neural network model for automatic recognition, and it is determined whether the sea surface image data collected in real time contains red tide areas and green tide areas, and the recognition results are output;

[0093] When it is determined that red tide areas and green tide areas appear in the sea surface image data collected in real time, the coverage areas of the red tide areas and green tide areas in the sea surface image data are extracted; otherwise, the sea surface image data is collected again in real time for identification and judgment;

[0094] By analyzing the formula:

[0095] The sea surface pollution rate G is obtained; where Red represents the coverage area of ​​the red tide area, Green represents the coverage area of ​​the green tide area, Area represents the total area of ​​the sea surface, and ε Red represents the impact factor of the red tide area, ε Green Represents the influencing factor of the green tide area.

[0096] Specifically, according to the collected and preprocessed sea surface image data, the red tide area and green tide area in the sea surface image data are annotated to ensure the accuracy and completeness of the annotation. The annotated image data is divided into a training set and a test set for subsequent model training and verification. The convolutional neural network model architecture is selected, including: ResNet, VGG, etc., and the model is trained using the annotated training set image data. The recognition performance of the model is improved by adjusting the model parameters and optimizing the algorithm. During the training process, the model performance is evaluated using the validation set to ensure that the model is not overfitting or underfitting. The sea surface image data is collected in real time through the marine ecological evaluation and analysis system based on big data. The real-time collected image data is preprocessed to match the input requirements of the trained convolutional neural network model. The preprocessed real-time sea surface image data is input into the trained convolutional neural network model for automatic recognition. The model outputs the recognition results, including the location, size, shape and other information of the red tide area and the green tide area. According to the recognition results of the model, the coverage area of ​​the red tide area and the green tide area in the sea surface image data is extracted. Image processing techniques such as image segmentation and morphological processing are used to accurately calculate the coverage area. The sea surface pollution rate is calculated according to the formula. The influence factor of the red tide area ε Red is 0.6, and the impact factor of the green tide area ε Green The influence factors can be set or adjusted according to the actual situation. The size and change trend of sea surface pollution rate are analyzed to evaluate the pollution status of the sea surface.

[0097] Feature extraction is performed based on the collected and processed marine remote sensing data, and preliminary productivity coefficients and coral reef bleaching change rates are calculated based on the extracted characteristic parameters, including the following steps:

[0098] Extract characteristic parameters from the collected and processed ocean remote sensing data by image segmentation and classification methods, wherein the characteristic parameters include: ocean surface temperature and chlorophyll concentration; the ocean remote sensing data includes: multispectral image data and hyperspectral image data;

[0099] Based on the collected and processed ocean remote sensing data, deep learning models are used to analyze and identify phytoplankton areas and coral reef bleaching areas in the ocean remote sensing data;

[0100] Label the hyperspectral image data of phytoplankton areas and coral reef bleaching areas, and input the labeled data into the deep learning model for training;

[0101] According to the trained deep learning model, the marine remote sensing data collected and processed in real time by the marine ecological evaluation and analysis system based on big data is input into the trained deep learning model to automatically identify the phytoplankton areas and coral reef bleaching areas and output the identification results;

[0102] According to the recognition result, the ocean remote sensing data collected historically at time point t is represented as a first remote sensing image, and the ocean remote sensing data collected in real time at time point t is represented as a second remote sensing image;

[0103] By analyzing the formula:

[0104] The preliminary productivity coefficient β is obtained; wherein A represents the area of ​​phytoplankton in the first remote sensing image, B represents the area of ​​phytoplankton in the second remote sensing image, N represents the chlorophyll concentration, η represents the conversion coefficient, η=3.7, and T represents the sea surface temperature;

[0105] By analyzing the formula:

[0106] Get the coral reef bleaching change rate C b ; Wherein, C1 represents the area of ​​coral reef bleaching in the first remote sensing image, and C2 represents the area of ​​coral reef bleaching in the second remote sensing image.

[0107] Specifically, the image segmentation technology is used to segment different areas in the marine remote sensing data. The characteristic parameters such as sea surface temperature and chlorophyll concentration are identified and extracted by classification methods. The phytoplankton area and coral reef bleaching area in the hyperspectral image data are manually labeled to ensure the accuracy and completeness of the annotation so as to be used for the subsequent training of the deep learning model. The labeled data is input into the deep learning model for training. The recognition performance and accuracy of the model are improved by adjusting the model parameters and optimizing the algorithm. According to the marine remote sensing data collected and processed in real time by the marine ecological evaluation and analysis system based on big data, the processed data is input into the trained deep learning model to automatically identify the phytoplankton area and the coral reef bleaching area, and the recognition result is output. According to the recognition result, the historical and real-time marine remote sensing data collected at the same time point t are obtained, which are represented as the first remote sensing image and the second remote sensing image respectively. The phytoplankton area is extracted from the first remote sensing image and the second remote sensing image. Combined with the chlorophyll concentration and the conversion coefficient, the preliminary productivity coefficient is calculated by the formula. This coefficient reflects the ability of phytoplankton in the marine ecosystem to convert solar energy into chemical energy through photosynthesis. Similarly, the coral reef bleaching area is extracted from the first remote sensing image and the second remote sensing image, and the coral reef bleaching change rate is calculated using a formula.

[0108] The comprehensive evaluation module comprises the following steps:

[0109] By analyzing the formula:

[0110]

[0111] The risk assessment index φ is obtained; among them, F represents the marine biological niche overlap index, E represents the impact rate of potential invasive species, R represents the marine water pollution rate, M represents the marine landform change rate, Y represents the marine resource change rate, G represents the sea surface pollution rate, β represents the initial productivity coefficient, C b represents the change rate of coral reef bleaching; W1, W2, W3, W4, W5, W6, W7 and W8 represent the weight coefficients of marine biological niche overlap index, potential invasive species impact rate, marine water quality pollution rate, marine geomorphological change rate, marine resource change rate, sea surface pollution rate, initial productivity coefficient and coral reef bleaching change rate, respectively;

[0112] Set thresholds and classify them according to the results of the risk assessment index to evaluate the risk level of marine ecology;

[0113] When φ<25%, the risk level of marine ecology is low risk;

[0114] When 25%≤φ≤60%, the risk level of marine ecology is medium risk;

[0115] When φ>60%, the risk level of marine ecology is high risk.

[0116] Specifically, relevant data such as the marine biological niche overlap index, the impact rate of potential invasive species, the marine water pollution rate, the marine geomorphology change rate, the marine resource change rate, the sea surface pollution rate, the initial productivity coefficient and the coral reef bleaching change rate are collected. According to expert opinions, historical data analysis and scientific research, the weight coefficients of various parameters are determined or adjusted, among which the weight coefficient of the marine biological niche overlap index is 0.1, the weight coefficient of the potential invasive species impact rate is 0.15, the weight coefficient of the marine water pollution rate is 0.1, the weight coefficient of the marine geomorphology change rate is 0.1, the weight coefficient of the marine resource change rate is 0.1, the weight coefficient of the sea surface pollution rate is 0.15, the weight coefficient of the initial productivity coefficient is 0.15, and the weight coefficient of the coral reef bleaching change rate is 0.15. The determination of the weight coefficient should take into account the degree and importance of the impact of each parameter on the marine ecological risk. The risk assessment index is calculated using a formula. This formula comprehensively considers multiple marine ecological parameters and their weight coefficients, and can comprehensively reflect the risk status of the marine ecology. According to the historical data and scientific analysis of the risk assessment index, the threshold is set. The setting of thresholds should be able to accurately distinguish different risk levels so that appropriate response measures can be taken. According to the results of the risk assessment index, it is compared with the set threshold to divide it into three levels: low risk, medium risk and high risk. When the risk assessment index is 25% lower than the set threshold, the risk level of the marine ecology is low risk; when the risk assessment index is between the 25% and 60% thresholds, it is medium risk; when the risk assessment index is higher than the highest threshold of 60%, it is high risk. Comprehensively analyze the risk assessment index and risk level to assess the overall risk status of the marine ecology.

Claims

1. A marine ecological evaluation and analysis system based on big data, characterized in that: Includes the following modules: Data acquisition module: collects marine ecological data, sea surface image data and marine remote sensing data; Data processing module: pre-processing the collected marine ecological data, including cleaning and standardization; pre-processing the collected sea surface image data, including denoising and image enhancement; pre-processing the collected marine remote sensing data, including denoising, geometric correction and radiation correction; Data analysis module: predict and analyze the distribution area of ​​marine biological species based on the collected and processed marine biological data and marine environmental data, and calculate the marine biological niche overlap index based on the prediction results; identify potential invasive species among marine biological species based on the predicted species distribution map, and calculate the impact rate of potential invasive species based on the analysis results; calculate the marine water quality pollution rate, marine landform change rate and marine resource change rate based on the marine ecological data analysis; identify and determine whether red tide areas and green tide areas appear based on the collected and processed sea surface image data, and calculate the sea surface pollution rate based on the identification results; extract features based on the collected and processed marine remote sensing data, and calculate the preliminary productivity coefficient and coral reef bleaching change rate based on the extracted feature parameters; The distribution areas of marine biological species are predicted and analyzed based on the collected and processed marine biological data and marine environmental data, and the marine biological niche overlap index is calculated based on the prediction results, including the following steps: Based on the collected and processed marine biological data and marine environmental data, the maximum entropy model is used to predict and analyze the distribution areas of marine biological species; Input the marine biological data and marine environmental data into the maximum entropy model for model training; according to the trained maximum entropy model, the marine ecological evaluation and analysis system based on big data collects and processes the marine biological data and marine environmental data in real time, inputs the processed marine biological data and marine environmental data into the trained maximum entropy model, predicts the distribution area of ​​marine biological species, and outputs the prediction results; according to the prediction results, converts the probability value of species occurrence in each grid unit into a species distribution map; According to the prediction results, by analyzing the formula: The marine biological niche overlap index F is obtained; where H represents the total number of marine resources, P ih represents the frequency of marine species i under marine resource h, P jh represents the frequency of marine biological species j under marine resource h; Comprehensive evaluation module: Calculate the risk assessment index of marine ecology based on comprehensive analysis of the marine biological niche overlap index, potential invasive species impact rate, marine water pollution rate, marine geomorphological change rate, marine resource change rate, sea surface pollution rate, initial productivity coefficient and coral reef bleaching change rate, and evaluate the marine ecology based on the analysis results; Feature extraction is performed based on the collected and processed marine remote sensing data, and preliminary productivity coefficients and coral reef bleaching change rates are calculated based on the extracted characteristic parameters, including the following steps: Extract characteristic parameters from the collected and processed ocean remote sensing data by image segmentation and classification methods, wherein the characteristic parameters include: ocean surface temperature and chlorophyll concentration; the ocean remote sensing data includes: multispectral image data and hyperspectral image data; Based on the collected and processed ocean remote sensing data, deep learning models are used to analyze and identify phytoplankton areas and coral reef bleaching areas in the ocean remote sensing data; Label the hyperspectral image data of phytoplankton areas and coral reef bleaching areas, and input the labeled data into the deep learning model for training; According to the trained deep learning model, the marine remote sensing data collected and processed in real time by the marine ecological evaluation and analysis system based on big data is input into the trained deep learning model to automatically identify the phytoplankton areas and coral reef bleaching areas and output the identification results; According to the recognition result, the ocean remote sensing data collected historically at time point t is represented as a first remote sensing image, and the ocean remote sensing data collected in real time at time point t is represented as a second remote sensing image; By analyzing the formula: The preliminary productivity coefficient β is obtained; wherein A represents the area of ​​phytoplankton in the first remote sensing image, B represents the area of ​​phytoplankton in the second remote sensing image, N represents the chlorophyll concentration, η represents the conversion coefficient, η=3.7, and T represents the sea surface temperature; By analyzing the formula: Get the coral reef bleaching change rate C b ; Wherein, C1 represents the area of ​​coral reef bleaching in the first remote sensing image, and C2 represents the area of ​​coral reef bleaching in the second remote sensing image; The comprehensive evaluation module comprises the following steps: By analyzing the formula: The risk assessment index φ is obtained; among them, F represents the marine biological niche overlap index, E represents the impact rate of potential invasive species, R represents the marine water pollution rate, M represents the marine landform change rate, Y represents the marine resource change rate, G represents the sea surface pollution rate, β represents the initial productivity coefficient, C b represents the change rate of coral reef bleaching; W1, W2, W3, W4, W5, W6, W7 and W8 represent the weight coefficients of marine biological niche overlap index, potential invasive species impact rate, marine water quality pollution rate, marine geomorphological change rate, marine resource change rate, sea surface pollution rate, initial productivity coefficient and coral reef bleaching change rate, respectively; Set thresholds and classify them according to the results of the risk assessment index to evaluate the risk level of marine ecology; When φ<25%, the risk level of marine ecology is low risk; When 25%≤φ≤60%, the risk level of marine ecology is medium risk; When φ>60%, the risk level of marine ecology is high risk.

2. The marine ecological evaluation and analysis system based on big data according to claim 1 is characterized in that: Collect marine ecological data, including: marine biological data, marine water quality data, marine resource data and marine environmental data; the marine biological data include: the number and type of marine biological species; the marine water quality data include: seawater salinity and pollutant concentration; the marine resource data include: fishery resources, mineral resources and renewable energy resources; the marine environmental data include: temperature, humidity, wind speed, dissolved oxygen and sediment quality.

3. The marine ecological evaluation and analysis system based on big data according to claim 1 is characterized in that: Identify potential invasive species among marine biological species based on predicted species distribution map analysis, and calculate the impact rate of potential invasive species based on the analysis results, including the following steps: Based on the predicted species distribution map, the predicted species distribution map is compared with the historical species distribution map, and species that do not appear in the historical species distribution map but appear in the predicted species distribution map are identified and marked as potential invasive species; By analyzing the formula: The impact rate E of potential invasive species is obtained; where S represents the total number of marine species and f represents the total number of potential invasive species.

4. The marine ecological evaluation and analysis system based on big data according to claim 1 is characterized in that: The calculation of ocean water pollution rate, ocean landform change rate and ocean resource change rate based on ocean ecological data analysis includes the following steps: By analyzing the formula: Get the ocean water pollution rate R; where C i represents the concentration of the i-th pollutant in the marine water quality data, Z i represents the standard limit of the concentration of the i-th pollutant, and m represents the total number of pollutants detected; The historical and real-time changes in sea level and marine resources at time point t; By analyzing the formula: The ocean landform change rate M is obtained; where ΔH represents the change in sea level height, and Δt represents the length of the time period; By analyzing the formula: The ocean resource change rate Y is obtained; wherein ΔQ represents the change in ocean resources, and Δt represents the length of the time period.

5. The marine ecological evaluation and analysis system based on big data according to claim 1 is characterized in that: According to the collected and processed sea surface image data, whether a red tide area and a green tide area appear is identified and judged, and the sea surface pollution rate is calculated according to the identification result analysis, including the following steps: Based on the collected and processed sea surface image data, the convolutional neural network model in the deep learning model is used to identify and determine whether red tide areas and green tide areas appear in the sea surface image data; Label the red tide areas and green tide areas in the sea surface image data; input the labeled image data into the convolutional neural network model for training; According to the trained convolutional neural network model, the sea surface image data collected and processed in real time by the marine ecological evaluation and analysis system based on big data is input into the trained convolutional neural network model for automatic recognition, and it is determined whether the sea surface image data collected in real time contains red tide areas and green tide areas, and the recognition results are output; When it is determined that red tide areas and green tide areas appear in the sea surface image data collected in real time, the coverage areas of the red tide areas and green tide areas in the sea surface image data are extracted; otherwise, the sea surface image data is collected again in real time for identification and judgment; By analyzing the formula: The sea surface pollution rate G is obtained; where Red represents the coverage area of ​​the red tide area, Green represents the coverage area of ​​the green tide area, Area represents the total area of ​​the sea surface, and ε Red represents the impact factor of the red tide area, ε Green Represents the influencing factor of the green tide area.

Images