A method and system for dividing giant panda activity areas based on big data

Abstract

The present invention discloses a method and system for dividing the activity area of ​​giant pandas based on big data, including: constructing a geographical weighted model according to the local characteristics of the giant panda's habitat space, then determining the potential spatial distribution according to the maximum entropy principle and known location and environmental data, simulating the influence of pheromones in the potential spatial distribution with an ant colony algorithm, deriving a predicted movement path through a hidden Markov chain, clustering the predicted movement path, using the Kriging interpolation method to obtain the activity intensity distribution and divide the area, and in each season, considering the impact of climate to generate a set of division strategies, and selecting the optimal division strategy for ecological protection and giant panda adaptability from the set as the final solution for the division of the giant panda's activity area. The method integrates multi-source data and multiple algorithms, improves the accuracy and scientificity of the division of the giant panda's activity area, helps its protection and ecological research, and has good interpretability.

Description

Technical Field

[0001] The present invention relates to the field of ecological environment science, and in particular to a method and system for dividing giant panda activity areas based on big data. Background Art

[0002] As a rare and endangered species in the world, effective protection and scientific management of giant pandas has become one of the important tasks of global biodiversity conservation. Accurately demarcating the activity areas of giant pandas is extremely important for gaining a deep understanding of their ecological habits, formulating reasonable protection strategies, and ensuring the integrity and connectivity of their habitats.

[0003] With the rapid development of modern information technology, big data is increasingly being used in the field of biodiversity research. In the study of giant pandas, a large amount of monitoring data has been accumulated, including location information of giant pandas obtained through GPS positioning technology, remote sensing images, habitat environmental data, and ecological data. However, traditional methods for dividing the activity areas of giant pandas often rely on limited field observations and simple statistical analysis, making it difficult to fully integrate and mine the complex information contained in multi-source big data. Moreover, these traditional methods usually ignore the impact of spatial heterogeneity of environmental factors on giant panda activities, as well as the dynamic characteristics of giant panda activities, such as seasonal migration and response to changes in food resources.

[0004] A method and system for dividing the activity areas of giant pandas based on big data, integrating big data and algorithm models, overcomes the shortcomings of traditional methods, and provides a more accurate, comprehensive and dynamically adaptive regional division plan for the protection of giant pandas, so as to better achieve effective protection and sustainable management of giant pandas and their habitats. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for dividing giant panda activity areas based on big data.

[0006] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0007] The first aspect of the present invention provides a method for dividing giant panda activity areas based on big data, comprising:

[0008] A geographically weighted model was constructed based on the local characteristics of different locations in the giant panda's habitat space;

[0009] Based on the maximum entropy principle and the geographically weighted model, the probability distribution that maximizes entropy is found according to the known location information and environmental data of giant pandas to obtain the potential spatial distribution of giant pandas;

[0010] In the potential spatial distribution of the giant panda, the influence of pheromones during the movement of the giant panda is simulated by an ant colony algorithm, and the predicted movement path of the giant panda is obtained based on a hidden Markov chain;

[0011] Clustering the predicted movement paths of the giant pandas, applying the Kriging interpolation method to each cluster to obtain the distribution of the giant panda activity intensity, and dividing them to obtain an activity area division strategy;

[0012] A set of partitioning strategies is obtained by considering the impact of climate in each season, and the partitioning strategy with the best ecological protection and giant panda adaptability is selected from the set of partitioning strategies for each season as the final partitioning method.

[0013] As a further method, the method of constructing a geographically weighted model based on the local features of different locations in the giant panda habitat space includes:

[0014] Collect satellite remote sensing images, habitat environment data and giant panda location data with timestamps, extract feature vectors after preprocessing, calculate the Pearson correlation coefficient between feature vectors and giant panda density, and use the top 20% feature vectors as local feature variables;

[0015] The Gaussian function is selected as the weight function of the geographical weighted model, and the location data of giant pandas is used as the dependent variable and the local characteristic variables are used as the independent variables to establish the geographical weighted model. The expression is:

[0016]

[0017] in, For the giant panda in position The intensity of activity, For location The coordinates of For location The constant term regression coefficient of is the number of independent variables, and The positions are No. The independent variable and independent variables, for The regression coefficient of For location No. The independent variable and The regression coefficients of the interactions of the independent variables are is the error term.

[0018] As a further method, the method of finding a probability distribution that maximizes entropy based on the maximum entropy principle combined with the geographical weighted model according to known giant panda location information and environmental data includes:

[0019] Based on the known location information and environmental data of giant pandas, a linear relationship between probability density and geographical weighted model is established through regression analysis, and the geographical weighted model is integrated into the entropy function. The expression is:

[0020]

[0021] in, is the probability distribution, The spatial area for the giant panda’s habitat. and are the intercept and slope terms of the linear function between the probability density function and the geographically weighted model, For geographically weighted models at location The value of

[0022] Determine constraints based on the ecological habits and biological characteristics of giant pandas, construct Lagrangian functions based on the constraints, find partial derivatives of the Lagrangian functions with respect to Lagrangian multipliers, and obtain a system of equations;

[0023] The system of equations is solved by the gradient descent method to obtain the probability distribution that satisfies the constraints and maximizes the entropy function as the potential space distribution of giant pandas.

[0024] As a further method, the method of simulating the influence of pheromones in the giant panda's movement process by using an ant colony algorithm in the potential spatial distribution of the giant panda includes:

[0025] Starting from a random starting position in the panda's latent space distribution, at each step, the probability that the ant will choose the next move position from its current position is expressed as:

[0026]

[0027] in, For the The ant moves from the current position to the next position The probability of From position To location The pheromone concentration, For location Heuristic information, and are the importance coefficients of pheromone concentration and heuristic information, For the Ants in position The collection of domain locations, From position To the field location The pheromone concentration, For field location Heuristic information;

[0028] The pheromone concentration on the path is updated, and the expression is:

[0029]

[0030] in, For time When, from the position To location The pheromone concentration of the path, is the pheromone evaporation coefficient, For in time When, from the position To location The pheromone concentration of the path, is the number of ants, For the Only ants are in the position To location The increase in pheromone released by the path;

[0031] The pheromone concentration on the path is iteratively updated until the average change rate of the ant's selected path is less than 0.01. The algorithm converges and the iteration is stopped to obtain the simulation results of the giant panda's movement trend.

[0032] As a further method, the method for obtaining the predicted movement path of the giant panda based on the hidden Markov chain includes:

[0033] The different positions of giant pandas in the latent space distribution are defined as the states of the hidden Markov chain, and the simulation results of the giant pandas' movement trends obtained by the ant colony algorithm are used as observations;

[0034] The probability of the giant panda being in each state at the initial moment is determined according to the potential space distribution of the giant panda, and the maximum path probability of each hidden state at each moment is calculated using the hidden Markov model. The expression is:

[0035]

[0036] in, For at the moment , along the path to the hidden state And the maximum probability of producing an observation sequence, For the initial moment Hidden The probability of In hidden state The first observation is observed The probability of is the total number of hidden states, For at the moment Reaching the hidden state The path, For at the moment Reaching the hidden state The set of all possible paths of For path The probability of occurrence, From the path At the moment Hidden state passed Move to hidden state The probability of In hidden state The next observed time Observed value The probability of

[0037] Based on the information of the previous hidden state of the optimal path at each moment, we backtrack to obtain the optimal hidden state at each moment and obtain the hidden state sequence, which is the predicted movement path of the giant panda.

[0038] As a further method, the method of clustering the predicted movement paths of the giant pandas, applying the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and dividing the clusters to obtain the activity area division strategy includes:

[0039] The predicted paths are analyzed using kernel density estimation to obtain spatial density. The paths with similarities greater than the preset threshold are divided into the same cluster. The variation function model is constructed based on the spatial characteristics of each cluster. The expression is:

[0040]

[0041] in, is the hysteresis moment, which indicates the spatial distance between two sampling points. The hysteresis moment is The number of point pairs is the number of sampling points. For the The sampling point locations, and Respectively at the location and The frequency of giant panda activities in the area;

[0042] Based on the variogram model, the Kriging interpolation method was used to interpolate each cluster to obtain the activity intensity value of each cluster. Based on the activity intensity, the giant panda habitat was divided into core activity areas, secondary activity areas, edge transition areas, potential activity areas and protection buffer areas. Specifically, the upper quartile of the activity intensity was calculated. , lower quartile and interquartile range , the activity intensity is greater than The area with activity intensity between and The area between is divided into secondary activity areas, and the activity intensity is divided into and The area between is divided into the edge transition zone, and the activity intensity is less than It is divided into potential activity areas, and the boundaries of the core activity areas are expanded outward by the average activity radius of giant pandas as a protection buffer zone.

[0043] As a further method, the method of obtaining a set of partitioning strategies by considering the climate impact for each season, and selecting the partitioning strategy with the best ecological protection and giant panda adaptability from the set of partitioning strategies for each season as the final partitioning method includes:

[0044] According to seasonal changes, the parameters of the variogram model are adjusted by considering the weights of different factors, and different activity intensity distributions are calculated to obtain a set of partitioning strategies;

[0045] The objective function of ecological protection and giant panda adaptability is constructed as follows:

[0046]

[0047] in, is the number of divided regions, For the divided into areas, For Region Neidi The ecological protection index value of each location, For Region Neidi The fitness index value of the giant panda at each location, For Region Neidi The food resource index value of each location, For Region Neidi The set of adjacent locations of a location, For Region Neidi Position and The distance between adjacent positions, For Region Neidi Position and The ecological similarity between adjacent locations;

[0048] The partitioning strategy with the highest objective function is selected from the partitioning strategy set as the final partitioning method.

[0049] The second aspect of the present invention provides a system for dividing giant panda activity areas based on big data, comprising:

[0050] The geographical weighting module is used to construct a geographical weighting model based on the local characteristics of different locations in the giant panda's habitat space;

[0051] A potential spatial distribution module is used to find a probability distribution that maximizes entropy based on the maximum entropy principle and the geographically weighted model according to known giant panda location information and environmental data to obtain the potential spatial distribution of giant pandas;

[0052] A movement simulation module, used to simulate the influence of pheromones during the movement of giant pandas by using an ant colony algorithm in the potential spatial distribution of the giant pandas, and to obtain a predicted movement path of the giant pandas based on a hidden Markov chain;

[0053] A strategy generation module is used to cluster the predicted movement paths of the giant pandas, apply the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and divide them to obtain the activity area division strategy;

[0054] The seasonal climate strategy module is used to obtain a set of partitioning strategies by considering the impact of climate in each season, and select the partitioning strategy with the best ecological protection and giant panda adaptability from the partitioning strategy set of each season as the final partitioning method.

[0055] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor; and a memory arranged to store computer executable instructions, wherein when the executable instructions are executed, the processor executes the method steps described in the first aspect.

[0056] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0057] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0058] (1) The present invention utilizes an ant colony algorithm to simulate the influence of pheromones and combines it with a hidden Markov chain to obtain a predicted movement path. This can capture the movement patterns and behavior models of giant pandas in their habitats, and can accurately predict the actual range of giant panda activities, thus avoiding the problem of too small or too large area division caused by only considering static positions.

[0059] (2) The present invention further refines the analysis of giant panda activities by clustering the predicted movement paths and applying the Kriging interpolation method to obtain the distribution of giant panda activity intensity. By understanding the activity intensity in different areas, a more hierarchical and accurate division of activity areas is achieved.

[0060] (3) The present invention considers the climate impact in each season and obtains a set of division strategies. By formulating corresponding division strategies according to the climate characteristics of different seasons, it can better adapt to the actual needs of giant pandas in different seasons and ensure that the divided activity areas are suitable for the survival and activities of giant pandas. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a flowchart of the steps of a method for dividing giant panda activity areas based on big data in an embodiment of the present invention.

[0062] Figure 2 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0064] Reference Figure 1 As shown, the present invention provides a method for dividing giant panda activity areas based on big data, comprising:

[0065] S100 builds a geographically weighted model based on the local characteristics of different locations in the giant panda’s habitat space;

[0066] It should be explained that the main purpose of constructing a geographically weighted model is to more accurately describe the relationship between the activity intensity of giant pandas at different locations in their habitat and various related factors. By considering the local characteristics of different locations, it is possible to capture the impact of the spatial heterogeneity of the habitat environment on the activities of giant pandas, rather than simply using a global unified model for analysis. This will make subsequent predictions and analyses of the distribution of giant panda activities more realistic;

[0067] In the actual evaluation, a nature reserve was selected as the research area, and Landsat satellite remote sensing images with a resolution of 30 meters in the past five years were collected. A total of 20 meteorological checkpoints and 30 soil monitoring points were set up in the reserve to collect environmental data including temperature and humidity, precipitation, soil type, and water source distribution. The data collection frequency was once an hour, and the location information of 15 giant pandas in the past five years was obtained. A total of about 44,000 valid location points were collected; a total of 50 eigenvectors were obtained, and the Pearson correlation coefficient between each eigenvector and the giant panda density was calculated. The eigenvectors in the top 20% were used as local feature variables to establish a geographically weighted model;

[0068] S200 searches for a probability distribution that maximizes entropy based on the maximum entropy principle and the geographical weighted model according to known giant panda location information and environmental data, and obtains the potential spatial distribution of giant pandas;

[0069] It needs to be explained that the maximum entropy principle is based on the ability to determine the probability distribution that best conforms to objective reality, has the greatest uncertainty and is free of subjective bias based only on existing known information. Combined with the geographically weighted model, the geographically weighted model integrates local characteristic variables such as altitude and vegetation coverage, and establishes a linear relationship with probability density through regression analysis, which can convert spatial characteristics into probability parameters. Considering the ecological habits and biological characteristics of giant pandas, such as their preference for specific altitudes and bamboo forest coverage and the requirements for concealment during the breeding season, as constraints, the Lagrangian function and gradient descent method are used to solve it, which can satisfy the natural behavior laws of giant pandas and infer their potential distribution probability at each location in the entire habitat space;

[0070] In the actual evaluation, the linear relationship between the probability density and the geographically weighted model was established through regression analysis based on the known location information and environmental data of 15 giant pandas, and the intercept term of the linear function was determined to be 0.1 and the slope term to be 0.5; the constraints were determined based on the ecological habits of giant pandas (preferring areas with an altitude of 1200-2500 meters and a bamboo forest coverage of more than 30%) and biological characteristics (high concealment requirements during the breeding period), and the Lagrangian function was constructed, the partial derivative was calculated, and the equation system was obtained, and the equation system was solved by the gradient descent method. After calculation, the probability distribution that satisfies the constraints and maximizes the entropy function is obtained, that is, the potential spatial distribution of giant pandas, including in an area of ​​about 20 square kilometers. After calculation, the potential distribution probability of giant pandas in different locations is between 0.05-0.7, among which the probability is the highest in the area with an altitude of 1800 meters, a bamboo vegetation coverage of 45% and close to valley streams (good concealment and sufficient water sources), reaching 0.7.

[0071] S300 simulates the influence of pheromones in the movement of giant pandas by using an ant colony algorithm in the potential spatial distribution of the giant pandas, and obtains a predicted movement path of the giant pandas based on a hidden Markov chain;

[0072] It needs to be explained that the ant colony algorithm can effectively capture this potential attraction factor and its impact on the movement path of giant pandas by simulating the dynamic changes of pheromones. The potential spatial distribution of giant pandas is divided into many nodes and paths, and the pheromone concentration reflects the relative attractiveness of each path. As the simulation progresses, the update mechanism of pheromone concentration simulates the giant panda's continuous exploration and feedback of different areas. In practice, if an area continues to provide rich resources, giant pandas will be more inclined to move around there and explore the surrounding areas. In this way, the ant colony algorithm provides an effective computational model for revealing the giant panda's movement decision-making mechanism in a complex habitat, which can provide a deeper insight into its behavior patterns, and thus provide a strong basis for protection and management work;

[0073] In the actual evaluation, the simulation started from a random starting position in the potential space distribution of giant pandas, and a total of 300 ants were set to participate. After 800 iterations, the algorithm converged and simulated the movement trends of giant pandas in different seasons and time periods, including: in summer, giant pandas tend to move to areas with higher altitudes, relatively low temperatures and vigorous bamboo growth; in winter, they tend to move to areas with lower altitudes, sheltered from the wind and close to valleys to avoid cold weather; the different positions of giant pandas in the potential space distribution are defined as the states of the hidden Markov chain, and the results of the ant colony algorithm simulation are used as observations. At the beginning of spring, the probability of giant pandas being in the bamboo area at an altitude of 1600-1900 meters is 0.35, the probability of being in the bamboo area at an altitude of 1900-2300 meters is 0.28, and the probability of being in other areas is 0.37. The hidden Markov model is used to calculate the maximum path probability of each hidden state at each moment, and the hidden state sequence, that is, the predicted movement path of the giant panda, is obtained. The results show that giant pandas will migrate between several major bamboo growth areas and will move around near water sources in the valley during specific time periods.

[0074] S400 clusters the predicted movement paths of the giant pandas, applies the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and divides them to obtain an activity area division strategy;

[0075] It needs to be explained that the Kriging interpolation method can calculate the activity intensity distribution of the entire region based on limited clustered sample points and taking into account spatial autocorrelation, and based on this, regional division can determine different levels of regions;

[0076] In the actual evaluation, kernel density estimation was used to perform spatial density analysis on the predicted paths, and those with similarity greater than 0.65 (preset threshold) were divided into the same cluster, resulting in a total of 5 clusters. A variogram model was constructed based on the spatial characteristics of each cluster, and the Kriging interpolation method was used to interpolate each cluster based on the variogram model to obtain the activity intensity value of each cluster, and the upper quartile of the activity intensity was calculated. , lower quartile , interquartile range , and regional divisions were carried out. The results for the spring division are as follows: the core activity area is about 25 square kilometers, the secondary activity area is about 30 square kilometers, the edge transition area is about 14 square kilometers, and the potential activity area is about 32 square kilometers. The boundary of the core activity area is expanded outward by the average activity radius of giant pandas (1.5 kilometers) as a protection buffer zone.

[0077] S500 considers the impact of climate in each season to obtain a set of partitioning strategies, and selects the partitioning strategy with the best ecological protection and giant panda adaptability from the set of partitioning strategies for each season as the final partitioning method.

[0078] In the actual evaluation, the parameters of the variogram model were adjusted according to seasonal changes. In spring, the weight of bamboo growth-related factors in the variogram model was increased; in summer, the weight of the impact of precipitation on water sources was increased; in autumn, the relevant parameters were adjusted considering the changes in the distribution of fruit food resources; in winter, the impact of temperature and wind shelter factors on the model was emphasized. The activity intensity distribution in different seasons was calculated by adjusting the parameters, and a set of partitioning strategies was obtained. The partitioning strategy with the highest objective function was selected from the partitioning strategy set as the final partitioning method. The summer partitioning results were as follows: Core activity area: concentrated in areas with an altitude of 1200-1600 meters, close to rivers, streams and other water-rich areas, with an area of ​​about 25 square kilometers; Secondary activity area: distributed outside the core activity area and adjacent hillside areas, with an altitude of 1000-1200 meters and 1600-1800 meters. meter interval, with an area of ​​about 35 square kilometers; marginal transition zone: involving areas with an altitude of 800-1000 meters and 1800-2000 meters, with an area of ​​about 23 square kilometers; potential activity area: including relevant areas with an altitude above 2000 meters and below 800 meters, with an area of ​​about 13 square kilometers; protection buffer zone: based on the boundary of the core activity zone, it expands outward by 2.5 kilometers to form a protection buffer zone.

[0079] In this embodiment, the method of constructing a geographically weighted model based on local features of different locations in the giant panda habitat space includes:

[0080] Collect satellite remote sensing images, habitat environment data and giant panda location data with timestamps, extract feature vectors after preprocessing, calculate the Pearson correlation coefficient between feature vectors and giant panda density, and use the top 20% feature vectors as local feature variables;

[0081] The Gaussian function is selected as the weight function of the geographical weighted model, and the location data of giant pandas is used as the dependent variable and the local characteristic variables are used as the independent variables to establish the geographical weighted model. The expression is:

[0082]

[0083] in, For the giant panda in position The intensity of activity, For location The coordinates of For location The constant term regression coefficient of is the number of independent variables, and The positions are No. The independent variable and independent variables, for The regression coefficient of For location No. The independent variable and The regression coefficients of the interactions of the independent variables are is the error term.

[0084] In this embodiment, the method of finding a probability distribution that maximizes entropy based on the maximum entropy principle and the geographical weighted model according to known giant panda location information and environmental data includes:

[0085] Based on the known location information and environmental data of giant pandas, a linear relationship between probability density and geographical weighted model is established through regression analysis, and the geographical weighted model is integrated into the entropy function. The expression is:

[0086]

[0087] in, is the probability distribution, The spatial area for the giant panda’s habitat. and are the intercept and slope terms of the linear function between the probability density function and the geographically weighted model, For geographically weighted models at location The value of

[0088] Determine constraints based on the ecological habits and biological characteristics of giant pandas, construct Lagrangian functions based on the constraints, find partial derivatives of the Lagrangian functions with respect to Lagrangian multipliers, and obtain a system of equations;

[0089] The system of equations is solved by the gradient descent method to obtain the probability distribution that satisfies the constraints and maximizes the entropy function as the potential space distribution of giant pandas.

[0090] In this embodiment, the method of simulating the influence of pheromones during the movement of giant pandas by using an ant colony algorithm in the potential spatial distribution of giant pandas includes:

[0091] Starting from a random starting position in the panda's latent space distribution, at each step, the probability that the ant will choose the next move position from its current position is expressed as:

[0092]

[0093] in, For the The ant moves from the current position to the next position The probability of From position To location The pheromone concentration, For location Heuristic information, and are the importance coefficients of pheromone concentration and heuristic information, For the Ants in position The collection of domain locations, From position To the field location The pheromone concentration, For field location Heuristic information;

[0094] The pheromone concentration on the path is updated, and the expression is:

[0095]

[0096] in, For time When, from the position To location The pheromone concentration of the path, is the pheromone evaporation coefficient, For in time When, from the position To location The pheromone concentration of the path, is the number of ants, For the Only ants are in the position To location The increase in pheromone released by the path;

[0097] The pheromone concentration on the path is iteratively updated until the average change rate of the ant's selected path is less than 0.01. The algorithm converges and the iteration is stopped to obtain the simulation results of the giant panda's movement trend.

[0098] In this embodiment, the method for obtaining the predicted movement path of the giant panda based on the hidden Markov chain includes:

[0099] The different positions of giant pandas in the latent space distribution are defined as the states of the hidden Markov chain, and the simulation results of the giant pandas' movement trends obtained by the ant colony algorithm are used as observations;

[0100] The probability of the giant panda being in each state at the initial moment is determined according to the potential space distribution of the giant panda, and the maximum path probability of each hidden state at each moment is calculated using the hidden Markov model. The expression is:

[0101]

[0102] in, For at the moment , along the path to the hidden state And the maximum probability of producing an observation sequence, For the initial moment Hidden The probability of In hidden state The first observation is observed The probability of is the total number of hidden states, For at the moment Reaching the hidden state The path, For at the moment Reaching the hidden state The set of all possible paths of For path The probability of occurrence, From the path At the moment Hidden state passed Move to hidden state The probability of In hidden state The next observed time Observed value The probability of

[0103] Based on the information of the previous hidden state of the optimal path at each moment, we backtrack to obtain the optimal hidden state at each moment and obtain the hidden state sequence, which is the predicted movement path of the giant panda.

[0104] In this embodiment, the method of clustering the predicted movement paths of the giant pandas, applying the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and dividing the clusters to obtain the activity area division strategy includes:

[0105] The predicted paths are analyzed using kernel density estimation to obtain spatial density. The paths with similarities greater than the preset threshold are divided into the same cluster. The variation function model is constructed based on the spatial characteristics of each cluster. The expression is:

[0106]

[0107] in, is the hysteresis moment, which indicates the spatial distance between two sampling points. The hysteresis moment is The number of point pairs is the number of sampling points. For the The sampling point locations, and Respectively at the location and The frequency of giant panda activities in the area;

[0108] Based on the variogram model, the Kriging interpolation method was used to interpolate each cluster to obtain the activity intensity value of each cluster. Based on the activity intensity, the giant panda habitat was divided into core activity areas, secondary activity areas, edge transition areas, potential activity areas and protection buffer areas. Specifically, the upper quartile of the activity intensity was calculated. , lower quartile and interquartile range , the activity intensity is greater than The area with activity intensity between and The area between is divided into secondary activity areas, and the activity intensity is divided into and The area between is divided into the edge transition zone, and the activity intensity is less than It is divided into potential activity areas, and the boundaries of the core activity areas are expanded outward by the average activity radius of giant pandas as a protection buffer zone.

[0109] In this embodiment, the method of obtaining a set of partitioning strategies by considering the climate impact for each season, and selecting the partitioning strategy with the best ecological protection and giant panda adaptability from the set of partitioning strategies for each season as the final partitioning method includes:

[0110] According to seasonal changes, the parameters of the variogram model are adjusted by considering the weights of different factors, and different activity intensity distributions are calculated to obtain a set of partitioning strategies;

[0111] The objective function of ecological protection and giant panda adaptability is constructed as follows:

[0112]

[0113] in, is the number of divided regions, For the divided into areas, For Region Neidi The ecological protection index value of each location, For Region Neidi The fitness index value of the giant panda at each location, For Region Neidi The food resource index value of each location, For Region Neidi The set of adjacent locations of a location, For Region Neidi Position and The distance between adjacent positions, For Region Neidi Position and The ecological similarity between adjacent locations;

[0114] The partitioning strategy with the highest objective function is selected from the partitioning strategy set as the final partitioning method.

[0115] The second aspect of the present invention provides a system for dividing giant panda activity areas based on big data, comprising:

[0116] The geographical weighting module is used to construct a geographical weighting model based on the local characteristics of different locations in the giant panda's habitat space;

[0117] A potential spatial distribution module is used to find a probability distribution that maximizes entropy based on the maximum entropy principle and the geographically weighted model according to known giant panda location information and environmental data to obtain the potential spatial distribution of giant pandas;

[0118] A movement simulation module, used to simulate the influence of pheromones during the movement of giant pandas by using an ant colony algorithm in the potential spatial distribution of the giant pandas, and to obtain a predicted movement path of the giant pandas based on a hidden Markov chain;

[0119] A strategy generation module is used to cluster the predicted movement paths of the giant pandas, apply the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and divide them to obtain the activity area division strategy;

[0120] The seasonal climate strategy module is used to obtain a set of partitioning strategies by considering the impact of climate in each season, and select the partitioning strategy with the best ecological protection and giant panda adaptability from the partitioning strategy set of each season as the final partitioning method.

[0121] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0122] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0123] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0124] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a device for dividing the activity area of ​​giant pandas based on big data at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the aforementioned methods for dividing the activity area of ​​giant pandas based on big data.

[0125] The present invention can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit, a network processor (Network Processor, NP), etc.; it can also be a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0126] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple applications, any one of the aforementioned giant panda activity area division system methods based on big data is executed.

[0127] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for dividing giant panda activity areas based on big data, characterized in that: The following steps are involved: A geographically weighted model is constructed according to the local features of different locations in the giant panda's habitat space; the method for constructing a geographically weighted model according to the local features of different locations in the giant panda's habitat space comprises: Collect satellite remote sensing images, habitat environment data and giant panda location data with timestamps, extract feature vectors after preprocessing, calculate the Pearson correlation coefficient between feature vectors and giant panda density, and use the top 20% feature vectors as local feature variables; The Gaussian function is selected as the weight function of the geographical weighted model, and the location data of giant pandas is used as the dependent variable and the local characteristic variables are used as the independent variables to establish the geographical weighted model. The expression is: in, For the giant panda in position The intensity of activity, For location The coordinates of For location The constant term regression coefficient of is the number of independent variables, and The location No. The independent variable and independent variables, for The regression coefficient of For location No. The independent variable and The regression coefficients of the interactions of the independent variables are is the error term; Based on the maximum entropy principle combined with the geographical weighted model, a probability distribution that maximizes entropy is found based on known giant panda location information and environmental data to obtain the potential spatial distribution of giant pandas; the method of finding a probability distribution that maximizes entropy based on the maximum entropy principle combined with the geographical weighted model and known giant panda location information and environmental data includes: Based on the known location information and environmental data of giant pandas, a linear relationship between probability density and geographical weighted model is established through regression analysis, and the geographical weighted model is integrated into the entropy function. The expression is: in, is the probability distribution, The spatial area for the giant panda’s habitat. and are the intercept and slope terms of the linear function between the probability density function and the geographically weighted model, For geographically weighted models at location The value of Determine constraints based on the ecological habits and biological characteristics of giant pandas, construct Lagrangian functions based on the constraints, find partial derivatives of the Lagrangian functions with respect to Lagrangian multipliers, and obtain a system of equations; The system of equations is solved by gradient descent method to obtain the probability distribution that satisfies the constraints and maximizes the entropy function as the potential space distribution of giant pandas; The ant colony algorithm is used to simulate the influence of pheromones in the movement of giant pandas in the potential space distribution of the giant pandas, and the predicted movement path of the giant pandas is obtained based on the hidden Markov chain; the method for obtaining the predicted movement path of the giant pandas based on the hidden Markov chain includes: The different positions of giant pandas in the latent space distribution are defined as the states of the hidden Markov chain, and the simulation results of the giant pandas' movement trends obtained by the ant colony algorithm are used as observations; The probability of the giant panda being in each state at the initial moment is determined according to the potential space distribution of the giant panda, and the maximum path probability of each hidden state at each moment is calculated using the hidden Markov model. The expression is: in, For at the moment , along the path to the hidden state And the maximum probability of producing an observation sequence, For the initial moment Hidden The probability of In hidden state The first observation is observed The probability of is the total number of hidden states, For at the moment Reaching the hidden state The path, For at the moment Reaching the hidden state The set of all possible paths of For path The probability of occurrence, From the path At the moment Hidden state passed Move to hidden state The probability of In hidden state The next observed time Observed value probability; Backtracking is performed based on the information of the previous hidden state of the optimal path at each moment to obtain the optimal hidden state at each moment, and the hidden state sequence is obtained, that is, the predicted movement path of the giant panda; Clustering the predicted movement paths of the giant pandas, applying the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and dividing them to obtain the activity area division strategy; the method of clustering the predicted movement paths of the giant pandas, applying the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and dividing them to obtain the activity area division strategy includes: The predicted paths are analyzed using kernel density estimation to obtain spatial density. The paths with similarities greater than the preset threshold are divided into the same cluster. The variation function model is constructed based on the spatial characteristics of each cluster. The expression is: in, is the hysteresis moment, which indicates the spatial distance between two sampling points. The hysteresis moment is The number of point logarithms when For the The sampling point locations, and Respectively at the location and The frequency of giant panda activities in the area; Based on the variogram model, the Kriging interpolation method was used to interpolate each cluster to obtain the activity intensity value of each cluster. Based on the activity intensity, the giant panda habitat was divided into core activity areas, secondary activity areas, edge transition areas, potential activity areas and protection buffer areas. Specifically, the upper quartile of the activity intensity was calculated. , lower quartile and interquartile range , the activity intensity is greater than The area with activity intensity between and The area between is divided into secondary activity areas, and the activity intensity is divided into and The area between is divided into the edge transition zone, and the activity intensity is less than Divide into potential activity areas, and expand the boundaries of the core activity area outward to the average activity radius of giant pandas as a protection buffer zone; Considering the impact of climate in each season to obtain a set of partition strategies, and selecting the partition strategy with the best ecological protection and giant panda adaptability from the partition strategy set of each season as the final partition method; the method of considering the impact of climate in each season to obtain a set of partition strategies, and selecting the partition strategy with the best ecological protection and giant panda adaptability from the partition strategy set of each season as the final partition method includes: According to seasonal changes, the parameters of the variogram model are adjusted by considering the weights of different factors, and different activity intensity distributions are calculated to obtain a set of partitioning strategies; The objective function of ecological protection and giant panda adaptability is constructed as follows: in, is the number of divided regions, For the divided into areas, For Region Neidi The ecological protection index value of each location, For Region Neidi The fitness index value of the giant panda at each location, For Region Neidi The food resource index value of each location, For Region Neidi The set of adjacent locations of a location, For Region Neidi Position and The distance between adjacent positions, For Region Neidi Position and The ecological similarity between adjacent locations; The partitioning strategy with the highest objective function is selected from the partitioning strategy set as the final partitioning method.

2. A method for dividing giant panda activity areas based on big data according to claim 1, characterized in that: The method of simulating the influence of pheromones in the giant panda's movement process by using an ant colony algorithm in the potential spatial distribution of the giant panda comprises: Starting from a random starting position in the panda's latent space distribution, at each step, the probability that the ant will choose the next move position from its current position is expressed as: in, For the An ant moves from its current position to the next position The probability of From position To location The pheromone concentration, For location Heuristic information, and are the importance coefficients of pheromone concentration and heuristic information, For the Ants in position The collection of domain locations, From position To the field location The pheromone concentration, For field location Heuristic information; The pheromone concentration on the path is updated, and the expression is: in, For time When, from the position To location The pheromone concentration of the path, is the pheromone evaporation coefficient, For in time When, from the position To location The pheromone concentration of the path, is the number of ants, For the Only ants are in the position To location The increase in pheromone released by the path; The pheromone concentration on the path is iteratively updated until the average change rate of the ant's selected path is less than 0.

01. The algorithm converges and the iteration is stopped to obtain the simulation results of the giant panda's movement trend.

3. A system for dividing giant panda activity areas based on big data, used to execute the method for dividing giant panda activity areas based on big data according to any one of claims 1 to 2, characterized in that: The system comprises: The geographical weighting module is used to construct a geographical weighting model based on the local characteristics of different locations in the giant panda's habitat space; A potential spatial distribution module is used to find a probability distribution that maximizes entropy based on the maximum entropy principle and the geographically weighted model according to known giant panda location information and environmental data to obtain the potential spatial distribution of giant pandas; A movement simulation module, used to simulate the influence of pheromones during the movement of giant pandas by using an ant colony algorithm in the potential spatial distribution of the giant pandas, and to obtain a predicted movement path of the giant pandas based on a hidden Markov chain; A strategy generation module is used to cluster the predicted movement paths of the giant pandas, apply the Kriging interpolation method to each cluster to obtain the distribution of the activity intensity of the giant pandas, and divide them to obtain the activity area division strategy; The seasonal climate strategy module is used to obtain a set of partitioning strategies by considering the impact of climate in each season, and select the partitioning strategy with the best ecological protection and giant panda adaptability from the partitioning strategy set of each season as the final partitioning method.

4. An electronic device, characterized in that: include: processor; as well as A memory arranged to store computer executable instructions, which, when executed, cause the processor to perform the method of any one of claims 1 to 2.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, which, when executed by an electronic device including a plurality of application programs, enable the electronic device to execute the method described in any one of claims 1 to 2.

Images