Water-power engineering fish resource loss compensation evaluation method

Abstract

The invention discloses a hydroelectric engineering fish resource loss compensation evaluation method, and relates to the technical field of ecological environment protection, and the method comprises the steps: dividing a target region of hydroelectric engineering construction, deploying a multi-source image collection device, synchronously obtaining a fish image and environment parameter data including flow velocity, water temperature and dissolved oxygen, and constructing an original database; and training a fish image recognition model based on a deep learning algorithm, and performing fish type recognition and quantity statistics in multiple scenes. According to the method, the image recognition technology and the digital twinborn technology are integrated, fish images and environment parameter data are synchronously acquired, a fish image recognition model is trained in combination with a deep learning algorithm, and automatic fish type recognition and real-time quantity statistics under multiple scenes are performed, so that fishes can be quickly and accurately recognized and counted; and the change trend of fish resources can be monitored in real time, the evaluation efficiency and accuracy are remarkably improved, and a scientific basis is provided for optimization of a resource scheduling strategy of compensation measures.

Description

Technical Field

[0001] This invention relates to the field of ecological and environmental protection technology, specifically to a method for assessing compensation for fish resource losses in hydropower projects. Background Technology

[0002] During the construction and operation of hydropower projects, they often have a significant impact on the surrounding ecological environment, especially aquatic biological resources. Fish resource loss is one of the more common and serious environmental impacts of hydropower project development. The construction of hydropower projects usually requires large-scale reservoir water storage, which will cause changes in the original water depth, water flow, temperature and other environmental conditions, affecting the habitat, growth and reproduction of fish.

[0003] Currently, the assessment of fish resource loss compensation in hydropower projects mainly relies on traditional methods such as ecological surveys, statistical analysis, and expert experience. However, traditional ecological survey methods depend on on-site sampling and monitoring, which are costly and time-consuming to obtain data and are difficult to reflect changes in fish resources in real time and simultaneously assess the effectiveness of compensation measures. Therefore, the problem to be solved by this invention is how to use image recognition technology to identify and count fish, and combine it with hydrological and water quality environmental parameters to construct a digital twin model of the hydropower project area, and then conduct dynamic simulation to predict the changing trend of fish resources and the effectiveness of compensation measures in order to optimize compensation measures. To this end, a method for assessing fish resource loss compensation in hydropower projects is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a method for assessing compensation for fish resource losses in hydropower projects, in order to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0006] A method for assessing compensation for fish resource loss in hydropower projects includes the following steps:

[0007] S1. Delineate the target area for hydropower project construction, deploy multi-source image acquisition equipment including underwater cameras and hydrological and water quality sensors, and simultaneously acquire fish images and environmental parameter data including flow velocity, water temperature and dissolved oxygen to build an original database;

[0008] S2. Train a fish image recognition model based on deep learning algorithm to automatically identify and count fish species in multiple scenarios, and generate a dynamic population distribution heat map.

[0009] S3. Analyze the coupling relationship between fish distribution and hydrological and water quality parameters, establish an environmental-biological response model, and quantify the impact weights of key factors on fish resources.

[0010] S4. Combining geographic information systems with environmental-biological response models, construct a virtual hydropower engineering area containing physical-biological-engineering elements based on digital twin technology;

[0011] S5. Simulate different ecological scheduling schemes in the digital twin platform to predict the changing trend of fish resources and the effect of compensation measures;

[0012] S6. Based on the simulation results, evaluate the effectiveness of fish resource loss compensation measures and generate a resource scheduling scheme for the optimal compensation measures.

[0013] A further improvement to the technical solution of the present invention is that: S1 includes:

[0014] Based on the geographical scope of the hydropower project and the distribution of fish habitats, the target area for the construction of the hydropower project is divided, and key areas including the reservoir area, the front area of ​​the dam, the downstream area of ​​the dam, and the tributary area are clearly defined.

[0015] Multiple monitoring stations were set up in each target area, and multi-source image acquisition equipment, including underwater cameras and hydrological and water quality sensors, were deployed to ensure that the equipment covered key areas. Among them, underwater cameras were used to capture images of fish, and hydrological and water quality sensors were deployed to monitor environmental parameters such as flow rate, water temperature and dissolved oxygen in real time.

[0016] The collected fish images and environmental parameter data are integrated, cleaned and organized to remove invalid and erroneous data, and then a raw database covering fish images and environmental parameters is constructed to store the fish image and environmental parameter data.

[0017] A further improvement to the technical solution of the present invention is that: S2 includes:

[0018] Fish image data covering multiple scenes were retrieved from public databases, the images were labeled to identify fish species, and the images were preprocessed, including noise reduction and contrast enhancement, to improve image quality and form the original dataset.

[0019] Feature analysis is performed on the original dataset to extract fish morphological features, including body outline, body proportion and body color. These features are then integrated to obtain a fish feature set, which is divided into a training set and a validation set.

[0020] A fish image recognition model was built using a deep learning algorithm framework based on convolutional neural networks. The model was trained using image data from the training set. By continuously adjusting the model parameters and evaluating the model performance using the validation set, the model structure was optimized to improve the model's recognition accuracy and generalization ability, thus obtaining a fully trained fish image recognition model.

[0021] The trained fish image recognition model is used to identify fish image data collected in real time, count the number of different fish species, and combine with geographic information system to generate dynamic population distribution heat map based on fish distribution location and quantity information, showing the distribution and activity of fish in different areas.

[0022] A further improvement to the technical solution of the present invention is that the process of fish image data identification by the fish image recognition model is as follows:

[0023] The trained fish image recognition model is deployed to the ecological monitoring system of the hydropower project to ensure that it can efficiently process the collected image data and receive real-time fish image data from multi-source image acquisition devices, covering fish activity in different locations in the hydropower project area. The fish image data is then converted in format and resized to adapt to the model input requirements. Fish morphological features, including body outline, body proportion and body color, are extracted from the data and input into the fish image recognition model for analysis.

[0024] The fish image recognition model compares and matches the fish's external features with known fish external features based on the fish feature patterns it has learned, calculates the feature similarity, and compares it with a preset similarity threshold to determine the specific species of fish in the image. After the recognition is completed, it automatically counts the number of different species of fish, and records relevant information such as recognition time and collection location, and generates a data report containing the species and number of fish.

[0025] The identified and statistically analyzed fish data are integrated into the ecological monitoring system of hydropower projects, and combined with geographic information and environmental parameter data to form an ecological monitoring dataset.

[0026] A further improvement to the technical solution of the present invention is that the calculation process of the feature similarity is as follows:

[0027] Based on a pre-prepared database containing known fish species and their physical characteristics, and for each fish physical characteristic, its weight is determined.

[0028] The fish shape features extracted from the input image are compared with the fish shape features in the known database. For each fish shape feature, the difference between the fish shape feature in the input image and the known fish shape features is calculated. The difference includes the difference in body shape outline, the difference in body proportion, and the difference in body surface color.

[0029] By combining the physical characteristics of each fish, the weighted difference degree is calculated, and the weighted difference degrees of all fish physical characteristics are added together to obtain the total weighted difference degree. Then, the feature similarity S is calculated by adding the reciprocal of the total weighted difference degree to 1.

[0030] For each known fish species, its feature similarity is compared with a preset similarity threshold θ. If S ≥ θ, the fish in the input image is considered to match the known fish species, and the specific species of the fish in the image is determined. If the feature similarity S of all known fish species is less than the similarity threshold θ, the fish in the input image are marked as unknown species.

[0031] A further improvement to the technical solution of the present invention is that: S3 includes:

[0032] Collect fish distribution data in hydropower project areas, covering fish species and quantity information in different regions and time periods, simultaneously acquire hydrological and water quality parameters, clean and standardize the data, remove outliers and missing values, and ensure the accuracy and consistency of the data;

[0033] Correlation analysis was used to analyze the coupling relationship between fish distribution and hydrological and water quality parameters, and key factors that significantly affect fish distribution were identified, including flow velocity, water temperature and dissolved oxygen.

[0034] Based on the analysis results, an environment-biological response model was constructed, the model structure and parameters were determined, the relationship between fish resources and the environment was simulated, and the impact weight of key factors on fish resources was quantified.

[0035] A further improvement to the technical solution of the present invention is that: S4 includes:

[0036] Collect geographic information data of the hydropower project area, and integrate the geographic information data with fish distribution data and hydrological and water quality parameters to form a geospatial database;

[0037] By combining an environmental-biological response model based on multiple linear regression with a geographic information system, the impact weights of key environmental factors on fish resources are quantified through model analysis, and these impact weights are mapped onto geospatial data.

[0038] By utilizing digital twin technology, geographic information systems and environmental-biological response models are integrated into a virtual hydropower project area to construct a virtual model that includes physical, biological, and engineering elements, enabling real-time dynamic simulation and visualization of the hydropower project area.

[0039] A further improvement to the technical solution of this invention lies in the following: the specific process of real-time dynamic simulation of the hydropower project area is as follows:

[0040] Different engineering construction scenarios are set up, namely normal operation scenario, engineering modification scenario, and extreme event scenario. The normal operation scenario simulates the situation of hydropower project under normal operating conditions, including reservoir water storage and release according to the design and scheduling plan, and hydropower station power generation at rated power. The engineering modification scenario simulates the ecosystem response under scenarios such as dam modification and the addition of ecological facilities. The extreme event scenario simulates the ecosystem response under extreme conditions such as floods and droughts.

[0041] For each engineering construction scenario, relevant parameters are set. During the simulation process, data of various elements in the region under different times and working conditions are periodically extracted from the digital twin model, including geographical environment data, water conservancy facility operation data and ecological data. The extracted data are sorted and classified to establish a database.

[0042] By accessing data from various elements in the database, we analyze the changes in fish species and numbers under different engineering construction scenarios, plot fish resource change curves, assess the impact of engineering construction on fish diversity and numbers, analyze the changing patterns of water flow velocity, water level, water temperature, and dissolved oxygen parameters under different scenarios, plot parameter change curves, assess the impact of engineering construction on water flow conditions and water quality, and then combine the fish resource change curves and parameter change curves to analyze the trend of the impact of engineering construction on fish resources and environmental parameters, and assess the changing trend of fish numbers under different scenarios.

[0043] A further improvement to the technical solution of the present invention is that: S5 includes:

[0044] Based on regional ecological characteristics, fish habits, and hydropower engineering conditions, various ecological scheduling schemes were designed, and the reservoir water storage and release strategies and flow settings under different schemes were clarified. These schemes were then accurately entered into the digital twin platform to ensure data integrity and accuracy.

[0045] Run various ecological scheduling schemes in the digital twin platform to simulate changes in the fish's living environment under different schemes in real time, predict the changing trends of fish resources in terms of quantity and species distribution, and generate visualized prediction results;

[0046] For different ecological scheduling schemes, corresponding compensation measures are set and input into the platform for simulation, and the differences in fish resource changes with and without compensation measures are compared.

[0047] A further improvement to the technical solution of the present invention is that: S6 includes:

[0048] Based on the results of digital twin simulation, we statistically analyzed the specific changes in fish resources in terms of quantity, species, and distribution range under different ecological scheduling schemes with and without compensation measures, and evaluated the protective effect of compensation measures on fish resources.

[0049] Based on the assessment results, the compensation measures were optimized and adjusted, and the implementation and scale of the compensation measures were improved. Based on the basic needs of hydropower project construction, the resource allocation scheme of the compensation measures that can effectively protect fish resources was selected as the optimal resource allocation scheme, so as to improve the protection effect of the compensation measures on fish resources.

[0050] Due to the adoption of the above technical solution, the technical progress achieved by this invention compared to the prior art is as follows:

[0051] 1. This invention provides a method for assessing fish resource loss compensation in hydropower projects. By integrating image recognition technology and digital twin technology, it utilizes the simultaneous acquisition of fish images and environmental parameter data, and combines deep learning algorithms to train a fish image recognition model to automatically identify fish species and count their numbers in real time under multiple scenarios. This method can not only quickly and accurately identify and count fish, but also monitor the changing trends of fish resources in real time, significantly improving the efficiency and accuracy of the assessment, and providing a scientific basis for optimizing resource scheduling strategies for compensation measures.

[0052] 2. This invention provides a method for assessing fish resource loss compensation in hydropower projects. Through digital twin simulation results, it can assess the changes in fish resources under different ecological scheduling schemes, and then make targeted optimizations and adjustments to compensation measures. This helps to improve the implementation of compensation measures, including adjusting the species, timing and quantity of fish released for stock enhancement, optimizing the layout and materials of artificial fish nests, etc., to ensure that resource input matches the protection effect, and can improve the accuracy and effectiveness of compensation measures, thereby more effectively protecting and restoring fish resources. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0054] Figure 1 This is a schematic diagram of the workflow of the present invention;

[0055] Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Example 1, such as Figure 1 , Figure 2 As shown, this invention provides a method for assessing compensation for fish resource losses in hydropower projects, comprising the following steps:

[0058] S1. Delineate the target area for hydropower project construction, deploy multi-source image acquisition equipment including underwater cameras and hydrological and water quality sensors, and simultaneously acquire fish images and environmental parameter data including flow velocity, water temperature, and dissolved oxygen to build a raw database. Based on the geographical scope of the hydropower project and the distribution of fish habitats, delineate the target area for hydropower project construction, and identify key areas including the reservoir area, the dam front area, the dam downstream area, and the tributary area. Set up multiple monitoring stations in each target area and deploy multi-source image acquisition equipment, including underwater cameras and hydrological and water quality sensors, to ensure that the equipment covers key areas. Among them, underwater cameras are used to capture fish images, and the deployed hydrological and water quality sensors monitor environmental parameter data such as flow velocity, water temperature, and dissolved oxygen in real time. The collected fish images and environmental parameter data are integrated, cleaned, and sorted to remove invalid and erroneous data, and then a raw database covering fish images and environmental parameters is built to store fish images and environmental parameter data.

[0059] The specific work involves: Based on the geographical scope of the hydropower project and the distribution of fish habitats, clearly defining the target areas for the hydropower project construction. These include key areas such as the reservoir area, the dam front area, the dam downstream area, and the tributary area. The reservoir area is the core area for water storage and level regulation; the dam front area is the section upstream of the dam, which is crucial for fish migration and habitat; the dam downstream area is the waters downstream of the dam, significantly affected by reservoir regulation; and the tributary area is the tributary region connected to the main river channel, providing diverse habitats for fish. Within each target area, multiple monitoring stations will be set up according to geographical features and ecological functions to ensure coverage of different regions and ecological environments. Multi-source image acquisition equipment will be deployed, including underwater cameras and hydrological and water quality sensors. Underwater cameras will be used to capture images of fish, ensuring coverage of fish activity. The system identifies key areas for fish movement to obtain real-time information on fish species and numbers. Simultaneously, it deploys hydrological and water quality sensors to monitor environmental parameters such as flow velocity, water temperature, and dissolved oxygen in real time, acquiring relevant information from multiple dimensions within the hydropower project area. The system integrates fish images and environmental parameter data collected from multiple sources, cleaning the raw data to remove invalid and erroneous data, ensuring accuracy and reliability. The cleaned data is then organized and labeled to give it clear meaning and a standardized format. Fish image data is categorized and labeled, clearly recording the shooting time and location information for each image. Environmental parameter data is organized according to time series and correlated with the corresponding fish image data, ultimately constructing a raw database encompassing both fish images and environmental parameters.

[0060] S2. A fish image recognition model is trained based on a deep learning algorithm to automatically identify and count fish species in multiple scenarios, generating a dynamic population distribution heatmap. Fish image data covering multiple scenarios is retrieved from public databases, the images are labeled, fish species are identified, and the images are preprocessed, including denoising and contrast enhancement, to improve image quality and form a raw dataset. Feature analysis is performed on the raw dataset to extract fish morphological features, including body contour, body proportions, and body color. This is integrated to obtain a fish feature set, which is then divided into training and validation sets. Specifically, image processing techniques are used to extract fish body contour features to describe the overall shape of the fish, and the proportions of different body parts are measured to reflect the fish's... The study investigated the growth and development characteristics and morphological differences of fish, recorded their body color information including color type, distribution, and brightness, and built a fish image recognition model using a deep learning algorithm framework based on convolutional neural networks. The model was trained using image data from the training set, and its performance was evaluated using the validation set by continuously adjusting the model parameters and optimizing the model structure to improve the model's recognition accuracy and generalization ability. The trained fish image recognition model was then used to identify fish image data collected in real time, count the number of different fish species, and, in conjunction with a geographic information system, generate a dynamic population distribution heat map based on the fish's distribution location and quantity information to show the distribution and activity of fish in different regions.

[0061] The specific work involves: extensively retrieving fish image data from public databases, covering various scenarios and including images of fish under different water quality, lighting, and aquatic environment conditions, to ensure data diversity and richness. The acquired images are meticulously annotated to identify the specific fish species in each image. Preprocessing is then performed on the annotated images, using denoising algorithms to remove noise interference and enhance image clarity. Contrast enhancement makes the distinction between fish and background more obvious, effectively improving image quality and ultimately forming the original dataset for model training. Feature analysis is then performed on the original dataset, extracting fish morphological features from the images, covering multiple dimensions such as body shape, proportions, and surface color. These features are systematically integrated to construct a fish feature set, which is then divided into training and validation sets. A fish image recognition model is built using a deep learning algorithm framework based on convolutional neural networks. The model is trained using the fish morphological features from the training set. During training, model parameters, including convolutional kernel size, number, and network parameters, are adjusted. The number of layers allows the model to gradually learn the characteristic patterns of fish. Simultaneously, the model's performance is evaluated using a validation set. Based on the evaluation results, the model structure is optimized to improve its recognition accuracy and generalization ability. After multiple iterations of training, a trained fish image recognition model is obtained. This trained model is then used to identify real-time collected fish image data, quickly determining the number of different fish species. Combining the spatial analysis and visualization functions of a geographic information system, a dynamic population distribution heatmap is generated based on the fish's distribution location and quantity information. This heatmap visually displays the distribution and activity of fish in different areas, clearly presenting the density and distribution trend of fish populations through color variations. In the heatmap, darker areas indicate a greater number of fish and a denser distribution, while lighter areas indicate a smaller number of fish and a sparser distribution.

[0062] Furthermore, the process by which the fish image recognition model identifies fish image data is as follows:

[0063] The trained fish image recognition model is deployed into the ecological monitoring system of the hydropower project to ensure its efficient processing of collected image data. It receives real-time fish image data from multi-source image acquisition devices, covering fish activity in different locations within the hydropower project area. The fish image data undergoes format conversion and size adjustment to suit the model's input requirements. Fish morphological features, including body outline, body proportions, and body color, are extracted and input into the fish image recognition model for analysis. Based on its learned fish feature patterns, the model compares and matches the fish morphological features with known fish morphological features, calculates feature similarity, and compares it with a preset similarity threshold to determine the specific species of fish in the image. After recognition, the model automatically counts the number of different fish species and records relevant information such as recognition time and collection location, generating a data report containing fish species and quantities. The identified and statistically analyzed fish data is then integrated into the ecological monitoring system of the hydropower project, combined with geographic information and environmental parameter data, to form an ecological monitoring dataset.

[0064] The process of calculating feature similarity is as follows:

[0065] Based on a pre-prepared database containing known fish species and their morphological features, each fish morphological feature includes body outline, body proportion, and body color. For each fish morphological feature, its weight is determined. The fish morphological features extracted from the input image are compared with the fish morphological features in the known database. For each fish morphological feature, the difference between the fish morphological feature in the input image and the known fish morphological features is calculated. The difference includes the difference in body outline, body proportion, and body color. Combining each fish morphological feature, a weighted difference is calculated. The weighted difference of all fish morphological features is summed to obtain the total weighted difference. Then, the reciprocal of the sum of weighted difference plus 1 is calculated to obtain the feature similarity S. For each known fish species, its feature similarity is compared with a preset similarity threshold θ. If S ≥ θ, the fish in the input image is considered to match the known fish species, and the specific species of the fish in the image is determined. If the feature similarity S of all known fish species is less than the similarity threshold θ, the fish in the input image is marked as an unknown species.

[0066] The expression for calculating feature similarity is as follows:

[0067]

[0068] In the formula, S represents the feature similarity, w represents the number of morphological features of the fish, and S represents the number of morphological features. i D represents the weight of the i-th fish morphological feature, used to adjust the importance of different features. iLet be the difference score of the i-th fish morphological feature, used to measure the difference between the fish to be identified and known fish species in this morphological feature. Let i be the index of the fish morphological feature, i = 1, 2, 3. The smaller the difference score, the higher the similarity. D1 is the body shape contour difference score. A is the total area of ​​the image, used for normalization. Ω is the domain of the image. C unknown (x,y) and C known (x, y) are the contour functions of the fish to be identified and the known fish species at position (x, y), respectively; D2 is the body proportion difference; L unknown and W unknown The length and width of the fish to be identified, L known and W known Given the length and width of a known fish species, D3 represents the body color difference, M is the total number of pixels in the image, c represents the color channels, and C... unknown,c (j) and C known,c (j) represents the color value (red, green, blue) of the c channel of the fish to be identified and the known fish species at the j-th pixel, respectively;

[0069] S3. Analyze the coupling relationship between fish distribution and hydrological and water quality parameters, establish an environmental-biological response model, quantify the impact weight of key factors on fish resources, collect fish distribution data in the hydropower project area, covering fish species and quantity information in different regions and time periods, simultaneously acquire hydrological and water quality parameters, clean and standardize the data, remove outliers and missing values ​​to ensure data accuracy and consistency, use correlation analysis to analyze the coupling relationship between fish distribution and hydrological and water quality parameters, identify key factors that significantly affect fish distribution, including flow velocity, water temperature and dissolved oxygen, based on the analysis results, construct an environmental-biological response model, determine the model structure and parameters, simulate the relationship between fish resources and the environment, and quantify the impact weight of key factors on fish resources;

[0070] The specific work involves: collecting comprehensive fish distribution data in the hydropower project area, covering fish species and quantities in different regions and at different times; simultaneously acquiring hydrological and water quality parameters for the area to gain a comprehensive understanding of the fish's habitat; and cleaning and standardizing the collected data to remove outliers and missing values, ensuring data accuracy and consistency. Data sources include field monitoring, automatic monitoring equipment, and historical data records. Correlation analysis methods are used to analyze the coupling relationship between fish distribution and hydrological and water quality parameters, identifying the degree of correlation between fish distribution and various hydrological and water quality parameters. Key factors significantly influencing fish distribution were identified as flow velocity, water temperature, and dissolved oxygen. The closer the absolute value of the correlation coefficient is to 1, the stronger the correlation; the closer it is to 0, the weaker the correlation. Both excessively high and low flow velocities are detrimental to fish survival and distribution. Excessive flow velocity increases swimming resistance, depletes fish energy, and may even wash them away from suitable habitats. Insufficient flow velocity can lead to eutrophication, affecting water quality and hindering fish growth and reproduction. Dams and sluice gates in hydropower projects alter the natural flow velocity of rivers. Different fish species exhibit varying tolerances to different water temperatures. In hydropower projects, suitable water temperatures promote feeding, digestion, and absorption in fish, enhancing their immunity and reproductive capacity. However, the impounding and releasing of water in reservoirs alters the local water temperature, thus affecting the habitat and activity range of fish. Fish absorb dissolved oxygen from the water through their gills for respiration to sustain life. When dissolved oxygen levels are too low, fish exhibit symptoms such as difficulty breathing, reduced activity, and loss of appetite, which can even lead to death in severe cases. Different fish species have varying dissolved oxygen requirements, and the operation of hydropower projects alters water flow and water layer structure, thereby affecting dissolved oxygen levels. Based on the results of correlation analysis, an environmental-biological response model was constructed using multiple linear regression as the basic framework. The identified key factors were used as independent variables, and the fish population was used as the dependent variable. The collected data were divided into training and test sets. The selected model was trained using the training set data, and the model parameters were adjusted. During the training process, cross-validation was used to prevent the model from overfitting. The trained model was evaluated using the test set data, and the model's prediction accuracy and mean squared error were calculated to evaluate the model's performance. Through model simulation, the impact weight of key factors on fish resources was quantified.

[0071] The formula for calculating the weights of key quantitative factors on fish resources is as follows:

[0072] N=β0+β1V+β2T+β3DO+∈;

[0073] In the formula, N represents the number of fish, which is the dependent variable of the model, indicating the number of fish observed under specific conditions; V represents the flow velocity, indicating the speed of water flow; T represents the water temperature, indicating the water temperature; DO represents the dissolved oxygen, indicating the dissolved oxygen content in the water; β0 is the intercept term, indicating the expected value of the number of fish when all independent variables are zero; β1 is the regression coefficient of flow velocity, indicating the expected change in the number of fish for each unit increase in flow velocity; β2 is the regression coefficient of water temperature, indicating the expected change in the number of fish for each unit increase in water temperature; and β3 is the regression coefficient of dissolved oxygen, indicating the expected change in the number of fish for each unit increase in dissolved oxygen. The expected change in quantity, ∈ represents the error term, indicating random errors or noise that the model cannot explain. Through a multiple linear regression model, the regression coefficients of each key factor are obtained. The larger the absolute value of the regression coefficient, the more significant the impact of that key factor on fish populations. If β1 is positive, it indicates that increased flow velocity will increase fish populations; if β1 is negative, it indicates that increased flow velocity will decrease fish populations. If β2 is positive, it indicates that increased water temperature will increase fish populations; if β2 is negative, it indicates that increased water temperature will decrease fish populations. If β3 is positive, it indicates that increased dissolved oxygen will increase fish populations; if β3 is negative, it indicates that increased dissolved oxygen will decrease fish populations.

[0074] S4. Combining geographic information systems with environmental-biological response models, construct a virtual hydropower engineering area containing physical-biological-engineering elements based on digital twin technology;

[0075] S5. Simulate different ecological scheduling schemes in the digital twin platform to predict the changing trend of fish resources and the effect of compensation measures;

[0076] S6. Based on the simulation results, evaluate the effectiveness of fish resource loss compensation measures and generate a resource scheduling scheme for the optimal compensation measures.

[0077] Example 2, as Figure 1 , Figure 2 As shown, based on Embodiment 1, the present invention provides a technical solution: preferably, S4 includes:

[0078] Geographic information data of the hydropower project area is collected and integrated with fish distribution data and hydrological and water quality parameters to form a geospatial database. An environmental-biological response model based on multiple linear regression is combined with the geographic information system. Through model analysis, the impact weights of key environmental factors on fish resources are quantified and mapped to geospatial data. Using digital twin technology, the geographic information system and the environmental-biological response model are integrated into the virtual hydropower project area to construct a virtual model containing physical, biological, and engineering elements, providing real-time dynamic simulation and visualization of the hydropower project area.

[0079] Furthermore, the specific process of real-time dynamic simulation of the hydropower project area is as follows:

[0080] Different engineering construction scenarios are set up, namely normal operation scenario, engineering modification scenario, and extreme event scenario. The normal operation scenario simulates the hydropower project under normal operating conditions, including reservoir impoundment and release according to the design scheduling plan, and hydropower station generating electricity at rated power. The engineering modification scenario simulates the ecosystem response under scenarios such as dam modification and the addition of ecological facilities. The extreme event scenario simulates the ecosystem response under extreme conditions such as floods and droughts. For each engineering construction scenario, relevant parameters are set. In the normal operation scenario, reservoir impoundment and release are carried out according to the design scheduling plan, hydropower station generates electricity at rated power, and the water level is controlled within the design range. The flow velocity is calculated based on the reservoir release flow. In the engineering modification scenario, the construction progress of dam modification and the addition of ecological facilities, as well as the inundation range parameters for surrounding land, are set. In the extreme event scenario, flood... The simulation process involves setting parameters for water flow, water level, and duration, as well as the duration of drought and water level decline. During the simulation, data on various elements within the region under different timeframes and operating conditions are periodically extracted from the digital twin model. This includes geographical environmental data, water conservancy facility operation data, and ecological data. The extracted data is then organized and categorized to establish a database. Data from each element in the database is retrieved to analyze changes in fish species and numbers under different engineering construction scenarios. Fish resource change curves are plotted to assess the impact of engineering construction on fish diversity and numbers. Furthermore, the variation patterns of water flow velocity, water level, water temperature, and dissolved oxygen parameters under different scenarios are analyzed, and parameter change curves are plotted to assess the impact of engineering construction on water flow conditions and water quality. Finally, by integrating the fish resource change curves and parameter change curves, the trend of the impact of engineering construction on fish resources and environmental parameters is analyzed, and the changing trend of fish numbers under different scenarios is assessed.

[0081] The specific work involves: collecting geographic information data of the hydropower project area, covering key elements such as topography, river course, and dam location; utilizing high-resolution satellite remote sensing imagery to acquire macro-geographic information such as topography and river course; extracting topographic feature parameters such as river width, curvature, and slope, as well as the location and outline information of water conservancy facilities such as dams and sluices, through image processing and analysis techniques; supplementing detailed information that satellite remote sensing cannot accurately obtain using field surveying; and simultaneously collecting fish distribution data and hydrological and water quality parameters, including information on fish species and quantities in different regions and time periods, as well as environmental parameters such as flow velocity, water temperature, and dissolved oxygen. The data is then integrated with geographic information data, fish distribution data, and hydrological and water quality parameters to construct a geospatial database. Within the hydropower project area, multiple fish monitoring stations are set up based on geographical features and ecological functions to ensure coverage of different regions and ecological environments. Underwater cameras are used to monitor and acquire information on fish species and quantities in different areas and time periods. Hydrological and water quality monitoring points are simultaneously set up at locations corresponding to the fish monitoring stations to ensure the temporal and spatial consistency of the acquired fish habitat data and fish distribution data. An environment-biological response model is constructed based on multiple linear regression, and the key links are quantified based on the regression coefficients of the multiple linear regression model. The influence weights of environmental factors on fish resources are determined by the magnitude of their regression coefficients. A larger absolute value indicates a more significant impact of the environmental factor on fish resources. This study combines an environmental-biological response model with a geographic information system (GIS), linking spatial data from a geospatial database with the influence weight data in the environmental-biological response model. Utilizing the spatial analysis capabilities of the GIS, the influence weights are mapped onto the geospatial data to represent the degree of influence of key environmental factors on fish resources at different geographical locations, achieving a dynamic correlation between environmental and biological data. Furthermore, digital twin technology is used to integrate the GIS and the environmental-biological response model into a virtual hydropower project area, constructing a virtual model. Including physical, biological, and engineering elements, the digital twin model is calibrated and verified by comparing it with monitoring data from actual hydropower engineering areas to ensure its accuracy and reliability. Model parameters are adjusted so that the virtual model can truly reflect the ecological and engineering characteristics of the actual engineering area. Through the digital twin model, the hydropower engineering area is simulated in real time, showing the changes of various elements in the area at different times and under different working conditions. Different engineering construction scenarios are set up to analyze the response and change trends of the hydropower engineering area's ecosystem under different scenarios. With the help of visualization technology, the simulation results of the digital twin model are presented in an intuitive graphical and image form.

[0082] S5 includes:

[0083] Based on regional ecological characteristics, fish habits, and hydropower engineering conditions, multiple ecological scheduling schemes were designed. The reservoir water storage and release strategies and flow settings under different schemes were clarified and accurately entered into the digital twin platform to ensure data integrity and accuracy. Each ecological scheduling scheme was run in the digital twin platform to simulate changes in the fish's living environment under different schemes in real time, predict the changing trends of fish resources in terms of quantity and species distribution, and generate visualized prediction results. For different ecological scheduling schemes, corresponding compensation measures were set and input into the platform for simulation, and the differences in fish resource changes with and without compensation measures were compared.

[0084] The specific work involves: comprehensively collecting relevant data on regional ecological characteristics, covering natural ecological elements such as topography and climate conditions; studying fish habits, including key ecological needs such as the breeding season, migration routes, suitable water temperature, and dissolved oxygen levels of different fish species; understanding hydropower project conditions, including reservoir capacity, regulation performance, power generation demand, and downstream irrigation and domestic water demand; designing multiple ecological scheduling schemes based on the survey results; planning reservoir water storage and release strategies for each scheme, clarifying the water storage and release flow rates, water level change amplitudes, and rates at different time periods; and comprehensively considering factors such as maintaining the basic functions of the river ecosystem and ensuring the water volume required for fish survival and reproduction, setting ecological flow rates to ensure that the ecological flow can meet the downstream ecological water demand; meticulously organizing the data in the designed ecological scheduling schemes; repeatedly calibrating the flow and water level data in the reservoir water storage and release strategies, as well as the ecological flow setting data; and inputting the organized and calibrated data into a digital twin platform to establish a data model corresponding to each ecological scheduling scheme. The system simulates reservoir operation and ecological flow changes under different schemes. Simulations of various ecological scheduling schemes are initiated within the digital twin platform. Based on the entered data, the system simulates the reservoir's water storage and release processes, as well as the dynamic changes in ecological flow under different schemes in real time. During the simulation, changes in the fish's living environment under different schemes are monitored and recorded in real time, including changes in parameters such as water temperature, dissolved oxygen, flow velocity, and water level. Using simulation results and data analysis functions, the system predicts the changing trends in fish resources in terms of quantity and species distribution under different ecological scheduling schemes, generating visualized prediction results. Corresponding compensation measures are formulated for different ecological scheduling schemes, including the construction of fish propagation and release stations, the setting up of artificial fish nests, and the improvement of fish habitat environments. The specific content, implementation time, and location information of the compensation measures are clarified. Relevant data on the compensation measures are input into the digital twin platform and compared with the scenario without compensation measures. The simulations simulate changes in fish resources under different ecological scheduling schemes with and without compensation measures, comparing the differences in fish resource changes with and without compensation measures.

[0085] S6 includes:

[0086] Based on the results of digital twin simulation, we statistically analyzed the specific changes in fish resources in terms of quantity, species, and distribution range under different ecological scheduling schemes with and without compensation measures. We evaluated the protective effect of compensation measures on fish resources, optimized and adjusted the compensation measures based on the evaluation results, improved the implementation and scale of the compensation measures, and selected the resource scheduling scheme with compensation measures that effectively protect fish resources as the optimal resource scheduling scheme based on the basic needs of hydropower project construction, so as to improve the protective effect of compensation measures on fish resources.

[0087] The specific work involves: conducting data statistics based on digital twin simulation results; for different ecological management schemes, compiling detailed data on changes in fish resources in three dimensions—quantity, species, and distribution range—with and without compensation measures. Specifically, for fish quantity, the data on increases and decreases at different time points under each scheme is analyzed; for species, the appearance and disappearance of fish species are recorded; and regarding distribution range, changes in fish habitats are identified. This process aims to assess the protective effect of compensation measures on fish resources and determine whether the measures have mitigated the adverse effects of ecological management on fish to a certain extent. Based on the assessment results, targeted optimization and adjustments are made to the compensation measures, analyzing data discrepancies and improving the implementation methods, including adjusting the species, timing, and quantity of fish released for stock enhancement. By optimizing the layout and materials of artificial fish nests and adjusting the scale of compensation measures, the aim is to ensure that resource input matches the protection effect, thereby improving the accuracy and effectiveness of compensation measures. Under the premise of meeting the basic needs of hydropower project construction, a comprehensive analysis of the protection effect of optimized compensation measures on fish resources is conducted. This includes a comprehensive evaluation of whether fish populations are steadily increasing, whether species are becoming more diverse, and whether the distribution range is being reasonably expanded. The overall protection effectiveness of fish resources under different schemes is compared, and the compensation measure that can maximize the effective protection of fish resources is selected. The resource allocation scheme for this compensation measure is then determined as the optimal resource allocation scheme, achieving coordinated development between hydropower project construction and fish resource protection, and improving the overall quality of compensation measures for fish resource protection.

[0088] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for assessing compensation for fish resource loss in hydropower projects, characterized in that, Includes the following steps: S1. Delineate the target area for hydropower project construction, deploy multi-source image acquisition equipment, and simultaneously acquire fish images and environmental parameter data including flow velocity, water temperature, and dissolved oxygen to build an original database. S2. Train a fish image recognition model based on deep learning algorithm, identify fish species and count their numbers in multiple scenarios, and generate a dynamic population distribution heat map. S3. Analyze the coupling relationship between fish distribution and hydrological and water quality parameters, establish an environmental-biological response model, and quantify the impact weights of key factors on fish resources. S4. Combining geographic information systems with environmental-biological response models, construct a virtual hydropower engineering area containing physical-biological-engineering elements based on digital twin technology; S5. Simulate different ecological scheduling schemes in the digital twin platform to predict the changing trend of fish resources and the effect of compensation measures; S6. Based on the simulation results, evaluate the effectiveness of fish resource loss compensation measures and generate a resource scheduling scheme for the optimal compensation measures.

2. The method for assessing compensation for fish resource loss in hydropower projects according to claim 1, characterized in that: S1 includes: Based on the geographical scope of the hydropower project and the distribution of fish habitats, the target area for the construction of the hydropower project is divided, and key areas including the reservoir area, the front area of ​​the dam, the downstream area of ​​the dam, and the tributary area are clearly defined. Multiple monitoring stations are set up in each target area, and multi-source image acquisition equipment is deployed, including underwater cameras and hydrological and water quality sensors. The underwater cameras are used to capture images of fish, and the deployed hydrological and water quality sensors are used to monitor environmental parameters such as flow rate, water temperature and dissolved oxygen in real time. The collected fish images and environmental parameter data are integrated, cleaned, and organized to construct a raw database covering fish images and environmental parameters for storing the data.

3. The method for assessing compensation for fish resource loss in hydropower projects according to claim 1, characterized in that: S2 includes: Fish image data covering multiple scenes were retrieved from public databases, the images were labeled, the fish species were identified, and the images were preprocessed to form the original dataset; Feature analysis is performed on the original dataset to extract fish morphological features, including body outline, body proportion and body color. These features are then integrated to obtain a fish feature set, which is divided into a training set and a validation set. A fish image recognition model is built using a deep learning algorithm framework based on convolutional neural networks. The model is trained using image data from the training set, and its performance is evaluated using the validation set. The model structure is then optimized to obtain a fully trained fish image recognition model. The trained fish image recognition model is used to identify fish image data collected in real time, count the number of different fish species, and combine with geographic information system to generate dynamic population distribution heat map based on fish distribution location and quantity information, showing the distribution and activity of fish in different areas.

4. The method for assessing compensation for fish resource loss in hydropower projects according to claim 3, characterized in that: The process by which the fish image recognition model identifies fish image data is as follows: The trained fish image recognition model was deployed to the ecological monitoring system of the hydropower project. Real-time fish image data was received from multi-source image acquisition equipment, covering fish activity in different locations in the hydropower project area. Fish morphological features, including body outline, body proportion and body color, were extracted from the data and input into the fish image recognition model for analysis. The fish image recognition model compares and matches the fish's external features with known fish external features based on the fish feature patterns it has learned, calculates the feature similarity, and compares it with a preset similarity threshold to determine the specific species of fish in the image. After the recognition is completed, it automatically counts the number of different species of fish, and records relevant information such as recognition time and collection location, and generates a data report containing the species and number of fish. The identified and statistically analyzed fish data are integrated into the ecological monitoring system of hydropower projects, and combined with geographic information and environmental parameter data to form an ecological monitoring dataset.

5. The method for assessing compensation for fish resource loss in hydropower projects according to claim 4, characterized in that: The calculation process for the feature similarity is as follows: Based on a pre-prepared database containing known fish species and their physical characteristics, and for each fish physical characteristic, its weight is determined. The fish shape features extracted from the input image are compared with the fish shape features in the known database. For each fish shape feature, the difference between the fish shape feature in the input image and the known fish shape features is calculated. The difference includes the difference in body shape outline, the difference in body proportion, and the difference in body surface color. By combining the physical characteristics of each fish, the weighted difference degree is calculated, and the weighted difference degrees of all fish physical characteristics are added together to obtain the total weighted difference degree. Then, the feature similarity S is calculated by adding the reciprocal of the total weighted difference degree to 1. For each known fish species, its feature similarity is compared with a preset similarity threshold θ. If S ≥ θ, the fish in the input image is considered to match the known fish species, and the specific species of the fish in the image is determined. If the feature similarity S of all known fish species is less than the similarity threshold θ, the fish in the input image are marked as unknown species.

6. The method for assessing compensation for fish resource loss in hydropower projects according to claim 1, characterized in that: S3 includes: Collect fish distribution data in hydropower project areas, covering fish species and quantities in different regions and time periods, and simultaneously acquire hydrological and water quality parameters, and clean and standardize the data. Correlation analysis was used to analyze the coupling relationship between fish distribution and hydrological and water quality parameters, and key factors that significantly affect fish distribution were identified, including flow velocity, water temperature and dissolved oxygen. Based on the analysis results, an environment-biological response model was constructed, the model structure and parameters were determined, the relationship between fish resources and the environment was simulated, and the impact weight of key factors on fish resources was quantified.

7. The method for assessing compensation for fish resource loss in hydropower projects according to claim 1, characterized in that: S4 includes: Collect geographic information data of the hydropower project area, and integrate the geographic information data with fish distribution data and hydrological and water quality parameters to form a geospatial database; By combining an environmental-biological response model based on multiple linear regression with a geographic information system, the impact weights of key environmental factors on fish resources are quantified through model analysis, and these impact weights are mapped onto geospatial data. By utilizing digital twin technology, geographic information systems and environmental-biological response models are integrated into a virtual hydropower project area to construct a virtual model that includes physical, biological, and engineering elements, enabling real-time dynamic simulation and visualization of the hydropower project area.

8. The method for assessing compensation for fish resource loss in hydropower projects according to claim 7, characterized in that: The specific process of real-time dynamic simulation of the hydropower project area is as follows: Different engineering construction scenarios are set up, namely normal operation scenario, engineering modification scenario, and extreme event scenario. The normal operation scenario simulates the situation of hydropower project under normal operating conditions; the engineering modification scenario simulates the ecosystem response under dam modification and new ecological facilities; and the extreme event scenario simulates the ecosystem response under extreme conditions of flood and drought. For each engineering construction scenario, relevant parameters are set. During the simulation process, data of various elements in the region under different times and working conditions are periodically extracted from the digital twin model, including geographical environment data, water conservancy facility operation data and ecological data. The extracted data are sorted and classified to establish a database. By accessing data from various elements in the database, we analyze the changes in fish species and numbers under different engineering construction scenarios, plot fish resource change curves, assess the impact of engineering construction on fish diversity and numbers, analyze the changing patterns of water flow velocity, water level, water temperature, and dissolved oxygen parameters under different scenarios, plot parameter change curves, assess the impact of engineering construction on water flow conditions and water quality, and then combine the fish resource change curves and parameter change curves to analyze the trend of the impact of engineering construction on fish resources and environmental parameters, and assess the changing trend of fish numbers under different scenarios.

9. The method for assessing compensation for fish resource loss in hydropower projects according to claim 1, characterized in that: S5 includes: Based on regional ecological characteristics, fish habits, and hydropower project conditions, various ecological scheduling schemes were designed, and the reservoir water storage and release strategies and flow settings under different schemes were clarified and accurately entered into the digital twin platform. Run various ecological scheduling schemes in the digital twin platform to simulate changes in the fish's living environment under different schemes in real time, predict the changing trends of fish resources in terms of quantity and species distribution, and generate visualized prediction results; For different ecological scheduling schemes, corresponding compensation measures are set and input into the platform for simulation, and the differences in fish resource changes with and without compensation measures are compared.

10. The method for assessing compensation for fish resource loss in hydropower projects according to claim 9, characterized in that: S6 includes: Based on the results of digital twin simulation, we statistically analyzed the specific changes in fish resources in terms of quantity, species, and distribution range under different ecological scheduling schemes with and without compensation measures, and evaluated the protective effect of compensation measures on fish resources. Based on the assessment results, the compensation measures were optimized and adjusted, and the implementation and scale of the compensation measures were improved. Based on the basic needs of hydropower project construction, the resource allocation scheme of the compensation measures that can effectively protect fish resources was selected as the optimal resource allocation scheme.

Images