Marine ecological protection and restoration project comprehensive evaluation method based on multi-dimensional analysis

Abstract

The invention discloses a marine ecological protection and restoration project comprehensive evaluation method based on multi-dimensional analysis, and relates to the field of marine ecological analysis, and the method comprises the steps: obtaining multi-dimensional monitoring data through constructing a target region map model and setting a monitoring point, and generating an ecological feature vector; calculating a relative distance and a deviation between the monitoring point and the comparison feature vector to obtain a repair expected goodness of fit; screening monitoring points with low goodness of fit and integrating the ecological data set; training and generating an optimized simulation feature vector based on a GAN model; and analyzing the simulation parameters and evaluating the ecological evolution trend to complete comprehensive evaluation. According to the method, multi-dimensional and high-precision ecological assessment is realized, the accuracy of restoration effect prediction and the scientificity of an assessment result are improved, reliable support is provided for a marine ecological protection restoration decision, and the problems of one-sided monitoring, insufficient prediction and the like of an existing method are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of marine ecological analysis, and more particularly, to a method for comprehensive evaluation of marine ecological protection and restoration projects based on multi-dimensional analysis. BACKGROUND

[0002] In recent years, the theory of marine ecological protection and restoration in China has been continuously improved and matured, and the work of marine ecological protection and restoration has also made some progress. However, there are still many problems, such as incomplete restoration system, non-uniform evaluation standards, etc. In particular, the scientificity, rationality and feasibility of the restoration scheme still have many blanks. Unscientific and unreasonable evaluation of the scheme directly leads to low adjustment ability of marine ecological protection and restoration projects, resulting in waste of manpower, material resources and funds. In addition, the existing technology lacks technical means for multi-dimensional analysis of the evolution trend of ecological characteristics between marine monitoring areas, making it difficult to comprehensively evaluate the scientificity and practicality of the marine restoration scheme. For marine monitoring platforms with limited monitoring resources and small amount of collected data, it is difficult to improve their comprehensive evaluation ability and achieve efficient and comprehensive restoration scheme evaluation process.

[0003] Therefore, there is an urgent need for a method for comprehensive evaluation of marine ecological protection and restoration projects based on multi-dimensional analysis. SUMMARY

[0004] The present application overcomes the defects of the prior art and provides a method for comprehensive evaluation of marine ecological protection and restoration projects based on multi-dimensional analysis.

[0005] The present application provides a method for comprehensive evaluation of marine ecological protection and restoration projects based on multi-dimensional analysis in the first aspect, comprising: S11: Constructing a map model according to the target marine monitoring area, setting a plurality of monitoring points, monitoring the monitoring points underwater through a monitoring unit, and obtaining remote sensing data of the monitoring points for comprehensive evaluation of ecological diversity, underwater biological activity distribution, vegetation condition, and water quality condition to obtain an ecological multi-dimensional feature vector; S12: Analyzing the expected ecological multi-dimensional parameters according to the plurality of monitoring points, and setting a comparison feature vector. For each monitoring point, the relative distance between the ecological multi-dimensional feature vector and the comparison feature vector is analyzed, and the relative distance deviation between the monitoring point and the adjacent monitoring point is calculated to obtain the restoration expected coincidence degree of each monitoring point; S13: Marking the monitoring points with a restoration expected coincidence degree less than the expected value as first monitoring points, and integrating the ecological multi-dimensional feature vectors of the first monitoring points and an adjacent monitoring point before and after ecological restoration to obtain an ecological data set; S14: Construct a GAN-based generation model to simulate training with the ecological data set as real data, obtain a simulated vector set, optimize the simulated vector set through a preset loss function and a judgment of the repair expected goodness of fit, and generate a simulated feature vector; S15: Perform ecological multi-dimensional parameter analysis on the simulated feature vector to generate a simulated ecological multi-dimensional parameter, and evaluate the marine ecological evolution trend of the monitoring point in the preset range based on the simulated ecological multi-dimensional parameter, and comprehensively score the marine repair.

[0006] In this scheme, S11 specifically comprises: Obtain the contour, range, and ecological vegetation distribution area information of the target marine monitoring area; Construct a three-dimensional visual map model according to the area information; Based on the ecological restoration planning scheme of the target marine monitoring area, set multiple monitoring points in the map model.

[0007] In this scheme, S11 further comprises: For each monitoring point, obtain high-resolution geographic remote sensing data, and based on an underwater monitoring unit, collect an image set of the monitoring point at different depths underwater; According to the remote sensing data, analyze the vegetation condition, introduce the NDVI and EVI parameter calculation, evaluate the density and coverage of the vegetation, and obtain the vegetation density and vegetation coverage; According to the image set, introduce an image recognition module to perform target recognition and statistics on underwater organisms, obtain the number of phytoplankton species and the number of underwater organism species, and perform activity trajectory statistical analysis on the target organisms within a unit time to calculate the activity distance and activity frequency of the underwater organisms; Obtain water quality parameters through the monitoring unit, including water temperature, turbidity, dissolved oxygen, pH value, ammonia nitrogen concentration, and nitrate concentration parameters; According to the ecological diversity, underwater organism activity distribution, vegetation condition, and water quality condition, set an ecological multi-dimensional feature vector.

[0008] In this scheme, S12 specifically comprises: According to the marine ecological condition and the repair expected state of the monitoring point, set an expected ecological multi-dimensional parameter for each monitoring point; Generate a comparison feature vector according to the expected ecological multi-dimensional parameter; According to each monitoring point, calculate the relative distance between the ecological multi-dimensional feature vector and the comparison feature vector, introduce Manhattan distance calculation in the calculation process, and obtain the relative distance of each monitoring point; Select one monitoring point as a repair monitoring point, calculate the average deviation of the relative distance between the repair monitoring point and the adjacent monitoring points, and set the repair expected goodness of fit of the repair monitoring point based on the average deviation. Set all monitoring points as repair monitoring points to calculate the repair expected fitness.

[0009] In this scheme, the S13 is specifically: Mark the monitoring point with a repair expected fitness less than the expected value as a first monitoring point. Obtain the ecological multi-dimensional feature vectors of the first monitoring point and one of the adjacent monitoring points before and after ecological restoration to obtain a first data set and a second data set. Integrate the first data set and the second data set to obtain an ecological data set.

[0010] In this scheme, the S14 is specifically: Construct a generative model based on GAN, which includes a generator G1 and a discriminator G2. Clean the ecological data set and import it into G1 as real data. In G1, learn the features of the real data, generate new simulated feature vectors, and optimize the model parameters through G2 to determine the authenticity of the simulated feature vectors. In the process of G1 and G2 adversarial training, introduce mean square error loss as the first loss, and distinguish the difference between simulated data and real data. The simulated feature vector is used as the ecological multi-dimensional feature vector of the first monitoring point to calculate the repair expected fitness, which is marked as the first fitness. Whether the first fitness is less than the expected value is used as the second loss, and is used for secondary discrimination of the authenticity of the simulated data. Perform cyclic adversarial training until the generated simulated data meets the first loss and the second loss. After cyclic adversarial training, a certain amount of simulated feature vectors are obtained through the generative model.

[0011] In this scheme, the S15 is specifically: Perform ecological multi-dimensional parameter analysis on the simulated feature vectors to generate simulated ecological multi-dimensional parameters. In the map model, mark the intermediate region based on the first monitoring point and one of the adjacent monitoring points. Use the simulated ecological multi-dimensional parameters as the ecological expected parameters of the intermediate region, and evaluate the marine ecological evolution trend of the monitoring points within the preset range. Analyze the marine ecological evolution trend of each monitoring point within the preset range, and comprehensively score the marine restoration project.

[0012] In this scheme, each monitoring point includes a corresponding ecological multi-dimensional feature vector and a comparison feature vector.

[0013] The second aspect of the present application also provides a comprehensive evaluation system for marine ecological protection and restoration projects based on multi-dimensional analysis, which comprises a memory, a processor and a data interface, wherein the memory comprises a comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis, and the comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis, when executed by the processor, implements the following steps: S11: constructing a map model according to a target marine monitoring area, setting a plurality of monitoring points, monitoring the monitoring points underwater through a monitoring unit, and obtaining remote sensing data of the monitoring points to comprehensively evaluate ecological diversity, underwater biological activity distribution, vegetation conditions and water quality conditions, and obtaining an ecological multi-dimensional feature vector; S12: analyzing expected ecological multi-dimensional parameters according to the plurality of monitoring points, and setting a comparison feature vector, for each monitoring point, analyzing the relative distance between the ecological multi-dimensional feature vector and the comparison feature vector, and calculating the relative distance deviation between the monitoring point and the adjacent monitoring point, to obtain a restoration expected coincidence degree of each monitoring point; S13: marking the monitoring points with a restoration expected coincidence degree less than an expected value as first monitoring points, integrating the ecological multi-dimensional feature vectors of the first monitoring points and an adjacent monitoring point before and after ecological restoration, to obtain an ecological data set; S14: constructing a generative model based on GAN, simulating training with the ecological data set as real data, obtaining a simulation vector set, optimizing and judging the simulation vector set through a preset loss function and a judgment of the restoration expected coincidence degree, and generating a simulation feature vector; S15: performing ecological multi-dimensional parameter analysis on the simulation feature vector, generating simulation ecological multi-dimensional parameters, and evaluating the marine ecological evolution trend of the monitoring points in a preset range based on the simulation ecological multi-dimensional parameters, and comprehensively scoring the marine restoration.

[0014] The third aspect of the present application also provides a computer readable storage medium comprising a comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis, and the comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis, when executed by a processor, implements the steps of the comprehensive evaluation method for marine ecological protection and restoration projects based on multi-dimensional analysis according to any one of the above aspects.

[0015] The application discloses a kind of based on multidimensional analysis's marine ecological protection restoration project comprehensive evaluation method, it is related to marine ecological analysis field, the method is by constructing target area map model and setting monitoring point, obtains multidimensional monitoring data and generates ecological characteristic vector;The relative distance and deviation of monitoring point and contrast characteristic vector are calculated, and the repair expected goodness of fit is obtained;Filtering low goodness of fit monitoring point and integrating ecological data set;Based on GAN model training generates optimized simulation characteristic vector;Analysis simulation parameter and evaluates ecological evolution trend, complete comprehensive evaluation.The application realizes multidimensional, high-precision ecological evaluation, improves the accuracy of repair effect prediction and the scientificity of evaluation result, provides reliable support for marine ecological protection restoration decision, solve the problem such as one-sidedness of existing method monitoring, prediction deficiency etc. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A flow chart of the marine ecological protection restoration project comprehensive evaluation method based on multidimensional analysis is shown; Figure 2 A multidimensional characteristic vector construction flow chart of the application is shown; Figure 3 A block diagram of the marine ecological protection restoration project comprehensive evaluation system based on multidimensional analysis is shown. DETAILED DESCRIPTION

[0017] In order to more clearly understand the above-mentioned purposes, features and advantages of the application, the application will be further described in detail below in combination with the drawings and specific embodiments. In the embodiments of the application, the embodiments of the application and the features in the embodiments can be combined with each other without conflict.

[0018] In the following description, many specific details are set forth in order to provide a thorough understanding of the application, but the application can also be implemented in other ways different from those described herein, therefore, the protection scope of the application is not limited by the specific embodiments disclosed below.

[0019] Figure 1 A flow chart of the marine ecological protection restoration project comprehensive evaluation method based on multidimensional analysis is shown.

[0020] As Figure 1 shown, the first aspect of the application provides a marine ecological protection restoration project comprehensive evaluation method based on multidimensional analysis, comprising: S11: constructing a map model according to a target marine monitoring area, setting a plurality of monitoring points, monitoring the monitoring points underwater through a monitoring unit, and obtaining remote sensing data of the monitoring points to comprehensively evaluate ecological diversity, underwater biological activity distribution, vegetation condition, and water quality, to obtain an ecological multidimensional characteristic vector; S12: According to the plurality of monitoring points, expected ecological multi-dimensional parameter analysis is carried out, and a comparison feature vector is set; for each monitoring point, relative distance analysis is carried out on the ecological multi-dimensional feature vector and the comparison feature vector, and the relative distance deviation of the monitoring point and the adjacent monitoring point is calculated, so as to obtain the repair expected fitting degree of each monitoring point; S13: The monitoring point with a repair expected fitting degree less than an expected value is marked as a first monitoring point; the ecological multi-dimensional feature vectors of the first monitoring point and an adjacent monitoring point before and after ecological restoration are integrated to obtain an ecological data set; S14: A generation model based on GAN is constructed, the ecological data set is used as real data for simulation training, a simulation vector set is obtained, the simulation vector set is optimized and judged through a preset loss function and the judgment of the repair expected fitting degree, and a simulation feature vector is generated; S15: The simulation feature vector is subjected to ecological multi-dimensional parameter analysis, simulation ecological multi-dimensional parameters are generated, and the monitoring point is evaluated based on the simulation ecological multi-dimensional parameters to evaluate the marine ecological evolution trend in a preset range, and a comprehensive score of marine restoration is generated.

[0021] According to the embodiment of the present application, the S11 is specifically: The contour, range and ecological vegetation distribution area information of the target marine monitoring area are obtained; A three-dimensional visual map model is constructed according to the area information; Based on the ecological restoration planning scheme of the target marine monitoring area, a plurality of monitoring points are set in the map model.

[0022] In the embodiment of the present application, the setting of the monitoring points is generally based on the average distribution of the target marine area, and the distance span between the monitoring points is generally large, so as to balance the demand for monitoring resources and the amount of monitoring data, and to minimize the redundant monitoring points, especially for the effective monitoring platform of resources.

[0023] Figure 2 The present application is shown in the ecological multi-dimensional feature vector construction flow chart, which is a simplified schematic diagram.

[0024] According to the embodiment of the present application, the S11 further includes: For each monitoring point, high-resolution geographic remote sensing data is obtained, and based on the underwater monitoring unit, an image set of the monitoring point at different depths underwater is collected; According to the remote sensing data, vegetation condition analysis is carried out, NDVI and EVI parameters are calculated, and the density and coverage of the vegetation are evaluated to obtain the vegetation density and vegetation coverage; According to the image set, an image recognition module is introduced to perform target recognition and statistics on the underwater organisms, to obtain the number of phytoplankton species and the number of underwater organism species, and to perform activity trajectory statistical analysis on the target organisms in a unit time, to calculate the activity distance and activity frequency of the underwater organisms. The water quality parameters are acquired by the monitoring unit, and the water quality parameters include water temperature, turbidity, dissolved oxygen, PH value, ammonia nitrogen concentration and nitrate concentration parameters. The ecological multi-dimensional feature vectors are set according to the ecological diversity, underwater organism activity distribution, vegetation condition and water quality condition.

[0025] In the embodiment of the application, the underwater monitoring unit includes an underwater camera device, a water quality detection device and a communication module. The underwater monitoring unit can simultaneously acquire water surface image information. The research on the water quality parameters is not limited to the above-mentioned parameters. Here, the vegetation density and vegetation coverage are used to reflect the vegetation condition, such as the mangrove forest and coastal wetland vegetation condition; the number of phytoplankton species and the number of underwater organism species are used to reflect the biological diversity; the activity distance and the activity frequency are used to reflect the underwater organism activity distribution; and the water quality parameters are used to reflect the water quality condition. All the above data are used as multi-dimensional ecological parameters and can be converted into feature vectors in the subsequent process. The analysis of the vegetation density can be based on the density evaluation of the number of vegetation sites within a certain range of the monitoring point. In the ecological multi-dimensional feature vectors, a plurality of (four) vectors can be generated based on a plurality of different categories of parameters (such as the four parameters of ecological diversity, underwater organism activity distribution, vegetation condition and water quality condition), or a multi-dimensional ecological feature vector can be generated based on all the parameters. In the subsequent vector comparative analysis, the vectors of the same dimension and the same parameter category are compared and analyzed.

[0026] In the activity trajectory statistical analysis of the target organisms in a unit time, the activity of the target underwater organisms in the core area of the monitoring point is mainly analyzed, and the activity trajectory is analyzed within a certain time, the activity frequency and the swimming distance are calculated, and the ecological activity degree is reflected.

[0027] According to the embodiment of the application, the S12, specifically: According to the marine ecological condition and the repair expected state of the monitoring point, the expected ecological multi-dimensional parameters of each monitoring point are set; The comparative feature vectors are generated according to the expected ecological multi-dimensional parameters; According to each monitoring point, the relative distance of the ecological multi-dimensional feature vectors and the comparative feature vectors is calculated, the Manhattan distance calculation is introduced in the calculation process, and the relative distance of each monitoring point is obtained; One monitoring point is marked as a repair monitoring point, the average deviation of the relative distance between the repair monitoring point and the adjacent monitoring point is calculated, and the repair expected coincidence degree of the repair monitoring point is set based on the average deviation. Set all monitoring points as repair monitoring points to calculate the repair expected fitness.

[0028] In the embodiment of the present application, the expected ecological multi-dimensional parameters include ecological diversity, underwater biological activity distribution, vegetation condition, water quality condition and the like. For example, for the water quality parameter, the water temperature can be set to 25°C, the turbidity to 4 NTU, the dissolved oxygen to 7 mg / L, the PH value to 7.8, the ammonia nitrogen concentration to 0.03 mg / L, and the nitrate concentration to 0.05 mg / L, and a feature vector is generated based on the multi-dimensional parameters. The comparative feature vector here is used to set an ideal state, and in the subsequent, a deviation analysis of the relative expected state (ecological dimension) is carried out based on the monitoring points and the adjacent points, and whether the ecological change trend conforms to the true trend is evaluated by the relative distance, and the analysis is carried out by using the related fitness parameters. The repair monitoring point is a monitoring point randomly selected from a plurality of monitoring points, or a point selected from the monitoring points according to a certain analysis sequence for evaluation. The expected ecological multi-dimensional parameters can be set according to the sea area where the monitoring point is located, the ecological condition, the vegetation state, the repair target and the like, and for the setting of the expected activity parameter, the information of the ecological diversity and the underwater biological activity distribution of an ideal sea area (or a successfully repaired marine area) can be referred to for setting, which has a certain relative reference significance. The relative distance reflects the relative deviation of the ecological characteristics of the monitoring point to the expected state.

[0029] The repair expected fitness is calculated as follows: ; Wherein, P represents the repair expected fitness of a certain repair monitoring point, N is the number of adjacent monitoring points for a certain repair monitoring point, are the relative distance of the repair monitoring point and the relative distance of the i-th adjacent monitoring point, respectively.

[0030] The adjacent monitoring point is a monitoring point adjacent in geographical position, or a monitoring point within a certain preset range with the repair monitoring point as the center.

[0031] It is worth noting here that the higher the fitness, the more the monitoring data within a certain range of the monitoring point conforms to the actual evolution trend, and through the fitness analysis, it can be evaluated which adjacent monitoring points deviate from the actual evolution trend or do not conform to the actual situation, and the higher the fitness, the more the ecological deterioration exists, and through the existing data of the monitoring points, it is difficult to analyze the ecological characteristic evolution trend of the intermediate area between the monitoring points and the monitoring points, and it is difficult to comprehensively evaluate the scientificity and practicality of the marine repair scheme. For the marine monitoring platform with limited monitoring resources and small amount of collected data, it is difficult to improve the comprehensive evaluation ability, and increasing the monitoring resources will undoubtedly greatly increase the consumption of manpower and material resources, and it is difficult to realize the efficient and intelligent marine repair comprehensive evaluation.

[0032] Based on this, the present application uses multi-dimensional ecological data to make relative expectation evaluation, analyzes the ecological expectation deviation of the monitoring point and the adjacent points, screens the adjacent monitoring points with low expectation coincidence degree, further generates vectorization and simulation features based on the related ecological data set of the adjacent monitoring points to predict the ecological evolution trend of the intermediate area between the actual adjacent monitoring points, further learns and simulates the monitoring point data set before and after the repair through the GAN model, uses the coincidence degree as the screening standard of the simulation data, generates the simulation ecological feature data for the intermediate area, and predicts and characterizes the ecological condition, so that the comprehensive evaluation of the marine repair is effectively realized, and a clear direction is provided for the subsequent repair scheme optimization.

[0033] Here, the ecological data set is generally the ecological data collected in multiple repair periods, and the ecological data before and after the ecological repair can be the ecological data before the repair to the repair stage, or the ecological data collected in the whole repair project and the whole stage.

[0034] Here, if the repair expectation coincidence degree is high, a more accurate value can be obtained in the form of linear evaluation or artificial evaluation according to the existing monitoring data, and if the coincidence degree is low, multi-dimensional feature generation is needed to evaluate the repair state and repair trend of the intermediate area to meet the accuracy analysis of the repair ecological condition.

[0035] According to the embodiment of the present application, the S13 is specifically: The monitoring point with a repair expectation coincidence degree less than the expected value is marked as a first monitoring point; Obtain the ecological multi-dimensional feature vectors of the first monitoring point and one of the adjacent monitoring points before and after the ecological repair to obtain a first data set and a second data set; Integrate the first data set and the second data set to obtain an ecological data set.

[0036] In the embodiment of the present application, the first data set and the second data set correspond to the first monitoring point and one of the adjacent monitoring points, and are both feature vector data. Here, between the first monitoring point and one of the adjacent monitoring points, there is a certain intermediate area between the monitoring points in geographical position, and in the subsequent simulation features, the ecological feature information of the intermediate area is simulated. The expected value is generally set to 50%.

[0037] According to the embodiment of the present application, the S14 is specifically: A generation model based on GAN is constructed, and the generation model includes a generator G1 and a discriminator G2; The ecological data set is data cleaned and imported into G1 as real data; In G1, the real data is feature learned, new simulation feature vectors are cyclically generated, the authenticity of the simulation feature vectors is judged through G2, and the model parameters are optimized; In the process of G1 and G2 confrontation training, the mean square error loss is introduced as the first loss, the difference between the simulated data and the real data is distinguished, the simulated feature vector is calculated as the ecological multi-dimensional feature vector of the first monitoring point, the repair expected fitness is calculated, and the first fitness is marked as the first fitness, whether the first fitness is less than the expected value is taken as the second loss, and is used for secondary discrimination of the authenticity of the simulated data; The cycle confrontation training is carried out until the generated simulated data meets the first loss and the second loss; After the cycle confrontation training, a certain amount of simulated feature vectors are obtained through the generation model.

[0038] In the embodiment of the application, the simulated feature vector generated by the generated model after the final training is used for subsequent analysis. The simulated feature vector generated in the intermediate training process is only used for training evaluation as simulated data. In addition to the first loss of mean square error, the improved generation model increases the repair expected fitness to distinguish whether the generated data meets a certain actual trend, and the second loss is used for model parameter update.

[0039] According to the embodiment of the application, the S15 is specifically: The simulated feature vector is subjected to ecological multi-dimensional parameter analysis to generate a simulated ecological multi-dimensional parameter; In the map model, the intermediate region is marked based on the first monitoring point and one of the adjacent monitoring points; The simulated ecological multi-dimensional parameter is taken as the ecological expected parameter of the intermediate region, and the marine ecological evolution trend of the monitoring point in the preset range is evaluated; The marine ecological evolution trend of each monitoring point in the preset range is analyzed, and the marine repair project is comprehensively scored.

[0040] In the embodiment of the application, the preset range includes the range formed by the first monitoring point, one of the adjacent monitoring points and the intermediate region.

[0041] Here, the first monitoring point can be analyzed with multiple adjacent monitoring points, and multiple intermediate regions are formed to perform comprehensive repair project evaluation, effectively expand the evaluation range, identify potential risks, improve the comprehensive evaluation capability of the overall region, and realize the in-process optimization capability in the repair project.

[0042] According to the embodiment of the application, each monitoring point includes a corresponding ecological multi-dimensional feature vector and a comparison feature vector.

[0043] It is worth mentioning that in the actual sea area monitoring platform, the evaluation of the repair project often needs to be based on the monitoring evaluation after the completion of the whole project, and it is difficult to predict the water quality change trend during the project, especially based on the small volume of the monitoring platform, the collected data are limited, and it is difficult to evaluate the potential risk and the rapid prediction of the marine ecological evolution trend, and it is difficult to evaluate which area has potential ecological fluctuation, so that only artificial experience judgment can be relied on, and the practicability and scientificity are not high. Based on this, the present application can improve the comprehensive evaluation ability of the marine ecological repair project by analyzing the expected coincidence degree of each monitoring point, mining the intermediate area with potential abnormal change, simulating the ecological characteristics through the generated model, and matching the actual ecological evolution trend.

[0044] According to the embodiment of the present application, further comprising: Based on the ecological restoration planning scheme, the repair route of the region is set, and the approach monitoring point is searched based on the repair route; The relative distance of the approach monitoring point is data sequenced, and the expected deviation sequence is obtained by sequentially referring to the route starting point sequence; The ARIMA algorithm is introduced, the autocorrelation and partial autocorrelation functions of the expected deviation sequence are calculated, and the p, d and q parameters are determined; The expected deviation sequence is predicted by the ARIMA algorithm, and the deviation trend of the ecological feature trend of the repair route and the expected repair state is evaluated according to the predicted data.

[0045] It can be understood that the repair route of the region here can be based on the repair route or the priority of the repair area of the repair scheme. The approach monitoring point can be searched based on the monitoring point within a certain radius range of the route. The expected deviation sequence is set for prediction analysis, which can judge the trend of the repair area route under the current repair state, and evaluate the direction of the overall repair effect.

[0046] P represents the autoregressive order of the time series, d represents the difference order of the time series, and q represents the moving average order of the time series.

[0047] According to the embodiment of the present application, the Figure 3 A block diagram of a marine ecological protection and restoration project comprehensive evaluation system based on multi-dimensional analysis is shown.

[0048] The second aspect of the present application also provides a comprehensive evaluation system for marine ecological protection and restoration projects based on multi-dimensional analysis, which comprises a memory 103, a processor 102 and a data interface 101, wherein the data interface is used for realizing data interaction function with a monitoring unit terminal 104, the memory comprises a comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis, and the comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis realizes the following steps when executed by the processor: S11: constructing a map model according to a target marine monitoring area, setting a plurality of monitoring points, monitoring the monitoring points underwater through a monitoring unit, and obtaining remote sensing data of the monitoring points to comprehensively evaluate ecological diversity, underwater biological activity distribution, vegetation condition and water quality condition, and obtaining an ecological multi-dimensional feature vector; S12: analyzing expected ecological multi-dimensional parameters according to a plurality of monitoring points, and setting a comparison feature vector, for each monitoring point, analyzing the relative distance between the ecological multi-dimensional feature vector and the comparison feature vector, and calculating the relative distance deviation between the monitoring point and the adjacent monitoring point, to obtain a restoration expected coincidence degree of each monitoring point; S13: marking the monitoring point with a restoration expected coincidence degree less than an expected value as a first monitoring point, integrating the ecological multi-dimensional feature vectors of the first monitoring point and an adjacent monitoring point before and after ecological restoration, to obtain an ecological data set; S14: constructing a generative model based on GAN, simulating training with the ecological data set as real data to obtain a simulation vector set, optimizing and judging the simulation vector set through a preset loss function and a judgment of the restoration expected coincidence degree, and generating a simulation feature vector; S15: analyzing the simulation feature vector in terms of ecological multi-dimensional parameters, generating simulation ecological multi-dimensional parameters, and evaluating the marine ecological evolution trend of the monitoring point in a preset range based on the simulation ecological multi-dimensional parameters, and comprehensively scoring the marine restoration.

[0049] The system can realize all steps of the above-mentioned comprehensive evaluation method for marine ecological protection and restoration projects based on multi-dimensional analysis according to monitoring requirements when in operation.

[0050] The third aspect of the present application also provides a computer readable storage medium comprising a comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis, and the comprehensive evaluation program for marine ecological protection and restoration projects based on multi-dimensional analysis realizes the steps of the comprehensive evaluation method for marine ecological protection and restoration projects based on multi-dimensional analysis according to any one of the above-mentioned aspects when executed by a processor.

[0051] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application can be generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (for example: coaxial cable, optical fiber, data user line) or wireless (for example: infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available media sets. The available medium can be a magnetic medium (for example: floppy disk, hard disk, magnetic tape), an optical medium (for example: digital versatile disc, or a semiconductor medium (for example: solid state disk) etc. In various embodiments of the present application, if there is no special description and logical conflict, the terms and / or descriptions of different embodiments are consistent and can be referred to each other, and the technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0052] In the embodiments of the present application, "at least one" means one or more, and "multiple" means two or more. The "and / or" describes the association relationship between the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the following three cases: A exists alone, A and B exist together, and B exists alone, where A and B can be singular or plural. In the textual description of the embodiments of the present application, the character " / ", generally indicates that the front and rear associated objects are in an "or" relationship. In the present application, "first", "second" and various number designations are only for the convenience of distinguishing description, and are not used to limit the scope of the embodiments of the present application. For example, to distinguish different messages, rather than to describe a specific order or sequence.

[0053] It can be understood that the various number designations involved in the embodiments of the present application are only for the convenience of distinguishing description, and are not used to limit the scope of the embodiments of the present application. The size of the serial number of the above processes does not mean the execution order, and the execution order of the processes should be determined by its function and inherent logic.

[0054] Finally, it should be noted that the above description is merely a specific implementation of the present application, and the protection scope of the present application is not limited thereto, and any change or replacement within the technical scope disclosed in the present application should be covered in the protection scope of the present application.

Claims

1. A method for comprehensive evaluation of marine ecological restoration projects based on multi-dimensional analysis, characterized in that, Comprise: S11: according to the target marine monitoring area, construct map model, set up a plurality of monitoring points, through the monitoring unit to the underwater monitoring of monitoring point, and obtain the remote sensing data of monitoring point to carry out comprehensive evaluation of ecological diversity, underwater biological activity distribution, vegetation condition, water quality, obtain the ecological multidimensional feature vector; S12: according to a plurality of monitoring points, the expected ecological multidimensional parameter analysis is carried out, and the contrast feature vector is set up, for each monitoring point, the relative distance analysis is carried out to the ecological multidimensional feature vector and the contrast feature vector, and the relative distance deviation of the monitoring point and the adjacent monitoring point is calculated, and the repair expected fitting degree of each monitoring point is obtained; S13: the monitoring point with repair expected fitting degree less than the expected value is marked as the first monitoring point, and the ecological multidimensional feature vectors of the first monitoring point and an adjacent monitoring point before and after the ecological restoration are integrated, and the ecological data set is obtained; S14: the generation model based on GAN is constructed, the ecological data set is used as the real data for simulation training, the simulation vector set is obtained, the simulation vector set is optimized and judged through the preset loss function and the judgment of repair expected fitting degree, and the simulation feature vector is generated; S15: the simulation feature vector is analyzed according to the ecological multidimensional parameter, the simulation ecological multidimensional parameter is generated, and the marine ecological evolution trend of the monitoring point in the preset range is evaluated based on the simulation ecological multidimensional parameter, and the comprehensive score of marine restoration is carried out.

2. The method according to claim 1, wherein, The S11 is specifically: obtain the contour, range, ecological vegetation distribution area information of the target marine monitoring area; construct a three-dimensional visual map model according to the area information; based on the ecological restoration planning scheme of the target marine monitoring area, a plurality of monitoring points are set in the map model.

3. The method according to claim 2, wherein, The S11 further comprises: for each monitoring point, high-resolution geographic remote sensing data is obtained, and based on the underwater monitoring unit, the image set of the monitoring point at different depths underwater is collected; according to the remote sensing data, the vegetation condition analysis is carried out, the NDVI and EVI parameters are introduced for calculation, the vegetation density and coverage rate are evaluated, and the vegetation density and vegetation coverage rate are obtained; according to the image set, the image recognition module is introduced to carry out target identification and statistics on underwater organisms, the number of phytoplankton species, the number of underwater organisms are obtained, the activity trajectory of target organisms in unit time is statistically analyzed, the activity distance and activity frequency of underwater organisms are calculated; the water quality parameters are obtained by the monitoring unit, including water temperature, turbidity, dissolved oxygen, PH value, ammonia nitrogen concentration and nitrate concentration parameters; according to the ecological diversity, underwater biological activity distribution, vegetation condition, water quality, the ecological multidimensional feature vector is set up.

4. The method according to claim 1, wherein, The S12 is specifically: according to the marine ecological condition and the repair expected state of the monitoring point, the expected ecological multidimensional parameter of each monitoring point is set up; the contrast feature vector is generated according to the expected ecological multidimensional parameter; according to each monitoring point, the relative distance calculation is carried out to the ecological multidimensional feature vector and the contrast feature vector, the Manhattan distance calculation is introduced in the calculation process, and the relative distance of each monitoring point is obtained; Screening out a monitoring point as a repair monitoring point, calculating the average deviation of the relative distance between the repair monitoring point and the adjacent monitoring points, and setting the repair expected fitness of the repair monitoring point based on the average deviation; Setting all monitoring points as repair monitoring points for repair expected fitness calculation.

5. The method according to claim 1, wherein, The S13, specifically: Marking the monitoring point with a repair expected fitness less than the expected value as a first monitoring point; Obtaining the ecological multi-dimensional feature vectors of the first monitoring point and one of the adjacent monitoring points before and after ecological restoration, obtaining a first data set and a second data set; Integrating the first data set and the second data set to obtain an ecological data set.

6. The method according to claim 5, wherein, The S14, specifically: Constructing a generative model based on GAN, the generative model including a generator G1 and a discriminator G2; Data cleaning the ecological data set and importing G1 as real data; In G1, feature learning is performed on the real data, new simulated feature vectors are generated in a loop, the authenticity of the simulated feature vectors is judged by G2, and the model parameters are optimized; In the process of G1 and G2 adversarial training, the mean square error loss is introduced as the first loss, the difference between the simulated data and the real data is judged, the simulated feature vectors are taken as the ecological multi-dimensional feature vectors of the first monitoring point to calculate the repair expected fitness, which is marked as the first fitness, and whether the first fitness is less than the expected value is taken as the second loss, which is used for secondary discrimination of the authenticity of the simulated data; Carrying out loop adversarial training until the generated simulated data meets the first loss and the second loss; After loop adversarial training, a certain amount of simulated feature vectors are obtained through the generative model.

7. The method according to claim 6, wherein, The S15 specifically: Performing ecological multi-dimensional parameter analysis on the simulated feature vectors to generate simulated ecological multi-dimensional parameters; In the map model, the intermediate region is marked based on the first monitoring point and one of the adjacent monitoring points; Taking the simulated ecological multi-dimensional parameters as the ecological expected parameters of the intermediate region, and evaluating the marine ecological evolution trend of the monitoring points within the preset range; Analyzing the marine ecological evolution trend of each monitoring point within the preset range, and comprehensively scoring the marine restoration project.

8. The method according to claim 1, wherein, Each monitoring point includes a corresponding ecological multi-dimensional feature vector and a comparison feature vector.

9. A multi-dimensional analysis-based comprehensive evaluation system for marine ecological restoration projects, characterized by, The system includes a memory, a processor, and a data interface, and the memory includes a multi-dimensional analysis-based marine ecological protection and restoration project comprehensive evaluation program, which is executed by the processor to implement the following steps: S11: Constructing a map model according to a target marine monitoring area, setting a plurality of monitoring points, monitoring the underwater monitoring points through a monitoring unit, and obtaining remote sensing data of the monitoring points for comprehensive evaluation of ecological diversity, underwater biological activity distribution, vegetation conditions, and water quality, to obtain an ecological multi-dimensional feature vector; S12: Perform expected ecological multi-dimensional parameter analysis on the plurality of monitoring points, and set a comparison feature vector. For each monitoring point, analyze the relative distance between the ecological multi-dimensional feature vector and the comparison feature vector, calculate the relative distance deviation between the monitoring point and the adjacent monitoring points, and obtain the repair expected fitness of each monitoring point; S13: Mark the monitoring point with a repair expected degree of coincidence less than the expected value as a first monitoring point, integrate the ecological multi-dimensional feature vectors of the first monitoring point and one adjacent monitoring point before and after ecological restoration to obtain an ecological data set; S14: Construct a generation model based on GAN, simulate training with the ecological data set as real data to obtain a simulation vector set, optimize and judge the simulation vector set through a preset loss function and a judgment of the repair expected degree of coincidence, and generate a simulation feature vector; S15: Perform ecological multi-dimensional parameter analysis on the simulation feature vector to generate a simulation ecological multi-dimensional parameter, and evaluate the marine ecological evolution trend of the monitoring point in a preset range based on the simulation ecological multi-dimensional parameter, and comprehensively score the marine restoration.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises a marine ecological protection and restoration project comprehensive evaluation program based on multi-dimensional analysis, and the marine ecological protection and restoration project comprehensive evaluation program based on multi-dimensional analysis, when executed by the processor, realizes the steps of the marine ecological protection and restoration project comprehensive evaluation method based on multi-dimensional analysis in any one of claims 1 to 8.

Images