Ecological problem identification method and device, storage medium and electronic device

Abstract

The present disclosure relates to the technical field of data processing, in particular to an ecological problem identification method and device, a storage medium and an electronic device. The ecological problem identification method comprises: discretizing a target region into a plurality of ecological units according to a grid, and constructing a multi-source ecological feature vector on the ecological units based on multi-source data of the target region; calculating initial ecological problem identification data on each ecological unit according to the multi-source ecological feature vector; discretizing the target region into a plurality of engineering units according to a constructable object of an ecological restoration technology, and constructing a two-part graph of the target region based on the ecological units and the engineering units; and converting the initial ecological problem identification data on each ecological unit into target ecological problem identification data on each engineering unit based on the two-part graph. The ecological problem identification method provided by the present disclosure can obtain ecological problem identification data which can be directly applied to ecological restoration technology implementation decisions.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, specifically to an ecological problem identification method, an ecological problem identification device, a storage medium, and an electronic device. Background Technology

[0002] Ecological problem identification is the cornerstone of ecological environmental protection and restoration. It uses scientific methods to discover, define, and analyze problems such as degradation, damage, or dysfunction in ecosystems.

[0003] Currently, the identification of ecological problems and the implementation of ecological restoration technologies often do not overlap spatially, resulting in a disconnect between "protection" and "engineering." Each system operates independently, information cannot be linked, data cannot be shared, and there is a lack of a unified data foundation and command and dispatch tools, which affects the effectiveness of governance.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this disclosure is to provide an ecological problem identification method, an ecological problem identification device, a storage medium, and an electronic device, aiming to solve the problem that ecological problem data cannot be directly applied to decision-making for the implementation of ecological restoration technologies.

[0006] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0007] According to one aspect of this disclosure, a method for identifying ecological problems is provided, comprising: The target area is discretized into multiple ecological units by grid, and a multi-source ecological feature vector is constructed on the ecological units based on the multi-source data of the target area; Initial ecological problem identification data are calculated for each ecological unit based on the multi-source ecological feature vector; The target area is discretized into multiple engineering units according to the constructable objects of ecological restoration technology, and a bipartite map of the target area is constructed based on the ecological units and the engineering units; Based on the bipartite graph, the initial ecological problem identification data in each of the ecological units will be converted into target ecological problem identification data in each of the engineering units.

[0008] Optionally, constructing a bipartite map of the target region based on the ecological units and the engineering units includes: An ecological unit set is constructed based on the ecological units, and an engineering unit set is constructed based on the engineering units; When the ecological unit and the engineering unit have a spatial relationship, an edge is established between the ecological unit and the engineering unit, and the edge weight of the edge is calculated; the spatial relationship includes one or more of the following: intersection, containment, adjacency, and connectivity; The bipartite graph is constructed based on the set of ecological units, the set of engineering units, the edges, and the edge weights.

[0009] Optionally, the step of converting the initial ecological problem identification data on each of the ecological units into target ecological problem identification data on each of the engineering units based on the bipartite graph includes: Based on the bipartite graph, edges including engineering units are determined to extract the corresponding ecological units and edge weights. For a given engineering unit, the weighted data of ecological problem identification on the corresponding ecological unit are summed to obtain the target ecological problem identification data on the engineering unit; wherein, the weighted data of ecological problem identification is the product of the initial ecological problem identification data on the ecological unit and the edge weight; The engineering units are traversed to obtain target ecological problem identification data for each engineering unit.

[0010] Optionally, the step of calculating the initial ecological problem identification data for each of the ecological units based on the multi-source ecological feature vectors includes: Design a joint loss function based on the data categories of the initial ecological problem identification data; The target deep learning network model is obtained by training the initial deep learning network model according to the joint loss function. The multi-source ecological feature vectors are input into the target deep learning network model to obtain the ecological problem identification data for each data category output by the target deep learning network model.

[0011] Optionally, the initial ecological problem identification data includes ecological problem probabilities and continuous ecological indicator data, and the design of a joint loss function based on the data type of the initial ecological problem identification data includes: A first loss function is constructed based on the class imbalance and hard samples of the semantic segmentation task, and the semantic segmentation task is used to calculate the probability of the ecological problem. The mean square error of the continuous ecological index data is used as the second loss function; The regularization loss term is used as the third loss function; the regularization loss term is either a consistent regularization loss term or a domain adaptive regularization loss term. The joint loss function is constructed based on the first loss function, the second loss function, and the third loss function.

[0012] Optionally, the method further includes: acquiring multi-source data of the target area, wherein the multi-source data includes one or more of optical image data, radar image data, and terrain data.

[0013] Optionally, the multi-source ecological feature vector constructed on the ecological unit based on multi-source data of the target region includes: Construct surface ecological parameters; the surface ecological parameters include one or more of the following: vegetation and land parameters, water body and water quality parameters, soil and salinity parameters, morphological and topographic parameters, and anthropogenic disturbance parameters; The parameter values ​​of the surface ecological parameters on the ecological unit are determined based on multi-source data; The multi-source ecological feature vector is constructed based on the parameter values.

[0014] According to a second aspect of this disclosure, an ecological problem identification device is provided, comprising: Vector unit, used to discretize the target area into multiple ecological units according to the grid, and to construct a multi-source ecological feature vector on the ecological unit based on the multi-source data of the target area; A computing unit is used to calculate initial ecological problem identification data on each of the ecological units based on the multi-source ecological feature vectors. The bipartite graph unit is used to discretize the target area into multiple engineering units according to the constructable objects of ecological restoration technology, and to construct a bipartite graph of the target area based on the ecological units and the engineering units; The conversion unit is used to convert the initial ecological problem identification data in each of the ecological units into target ecological problem identification data in each of the engineering units based on the bipartite graph.

[0015] According to a third aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the ecological problem identification method as described in the above embodiments.

[0016] According to a fourth aspect of the present disclosure, an electronic device is provided, characterized in that it includes: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the ecological problem identification method as described in the above embodiments.

[0017] The exemplary embodiments disclosed herein may have some or all of the following beneficial effects: In some embodiments of this disclosure, the technical solutions provide initial ecological problem identification data by constructing multi-source ecological feature vectors from multi-source data of the target area. Then, by constructing a bipartite graph of two-layer units—ecological units and engineering units—discrete the target area, the initial ecological problem identification data in each ecological unit is transformed into target ecological problem identification data in each engineering unit. On the one hand, using regular rasterized ecological units to calculate the initial ecological problem identification data and constructing standardized feature vectors can ensure the uniformity of data units, reduce invalid calculations, and significantly improve the efficiency of large-scale data processing. On the other hand, by pre-constructing a bipartite graph of two-layer units of ecological units and engineering units in the target area, the correspondence between the two types of heterogeneous nodes in ecological units and engineering units can be decoupled from the ecological problem identification calculation, enhancing the flexibility and maintainability of ecological problem identification. Furthermore, the identification data on the ecological units is transformed and accurately mapped to the engineering units through the constructed bipartite graph relationship network, transforming the unitized ecological problem distribution into decision-making information that can be directly used by engineering managers, thereby promoting the efficient implementation of ecological protection and restoration work.

[0018] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings: Figure 1 The illustration schematically shows a flowchart of an ecological problem identification method in an exemplary embodiment of the present disclosure; Figure 2 The illustration schematically shows a flowchart of a method for calculating initial ecological problem identification data in an exemplary embodiment of the present disclosure; Figure 3 The illustration schematically shows a flowchart of a bipartite graph construction method in an exemplary embodiment of the present disclosure; Figure 4 This schematic diagram illustrates the composition of an ecological problem identification device according to an exemplary embodiment of the present disclosure; Figure 5 The schematic diagram illustrates the structure of a computer system of an electronic device according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0020] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0021] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0022] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0023] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0024] The implementation details of the technical solutions of the embodiments of this disclosure are described in detail below.

[0025] Figure 1 This illustration schematically depicts a flowchart of an ecological problem identification method according to an exemplary embodiment of this disclosure. Figure 1 As shown, the ecological problem identification method includes steps S101 to S107: Step S101: Discretize the target area into multiple ecological units according to the grid, and construct a multi-source ecological feature vector on the ecological units based on the multi-source data of the target area; Step S103: Calculate the initial ecological problem identification data for each ecological unit based on the multi-source ecological feature vector; Step S105: Discretize the target area into multiple engineering units according to the constructable objects of ecological restoration technology, and construct a bipartite map of the target area based on the ecological units and the engineering units; Step S107: Based on the bipartite graph, the initial ecological problem identification data in each of the ecological units is converted into target ecological problem identification data in each of the engineering units.

[0026] In some embodiments of this disclosure, the technical solutions provide initial ecological problem identification data by constructing multi-source ecological feature vectors from multi-source data of the target area. Then, by constructing a bipartite graph of two-layer units—ecological units and engineering units—discrete the target area, the initial ecological problem identification data in each ecological unit is transformed into target ecological problem identification data in each engineering unit. On the one hand, using regular rasterized ecological units to calculate the initial ecological problem identification data and constructing standardized feature vectors can ensure the uniformity of data units, reduce invalid calculations, and significantly improve the efficiency of large-scale data processing. On the other hand, by pre-constructing a bipartite graph of two-layer units of ecological units and engineering units in the target area, the correspondence between the two types of heterogeneous nodes in ecological units and engineering units can be decoupled from the ecological problem identification calculation, enhancing the flexibility and maintainability of ecological problem identification. Furthermore, the identification data on the ecological units is transformed and accurately mapped to the engineering units through the constructed bipartite graph relationship network, transforming the unitized ecological problem distribution into decision-making information that can be directly used by engineering managers, thereby promoting the efficient implementation of ecological protection and restoration work.

[0027] The following will describe in more detail each step of the ecological problem identification method in this exemplary embodiment, with reference to the accompanying drawings and embodiments.

[0028] In step S101, the target area is discretized into multiple ecological units according to the grid, and a multi-source ecological feature vector is constructed on the ecological units based on the multi-source data of the target area.

[0029] Specifically, ecological units are regional units that divide the target area from an ecological perspective. For example, ecological units can be grids or patches, thereby better representing the ecological situation.

[0030] In one embodiment of this disclosure, the method further includes: acquiring multi-source data of the target area, wherein the multi-source data includes one or more of optical image data, radar image data, and terrain data.

[0031] Specifically, this involves collecting multi-temporal, multi-source data of the target area, including optical imagery, radar imagery, and topographic data. Optical imagery data includes Sentinel-2 (10-20m), Landsat-8 / 9 (15-30m), and high-resolution series data, or UAV aerial survey data, including visible light, multispectral, and thermal infrared imagery; radar imagery data, i.e., SAR (Synthetic Aperture Radar) data, such as Sentinel-1 or GF-3, can enhance the identification of water bodies, surface roughness, and shadow areas; topographic data includes DEM (Digital Elevation Model) data, including the topography, aspect, and slope of the target area.

[0032] In one embodiment of this disclosure, after collecting multi-source data, it can be preprocessed, such as radiometric calibration, atmospheric correction, geometric correction, noise suppression, cloud or cloud shadow removal and temporal synthesis, and finally uniformly resampled to the ecological unit.

[0033] Then, based on the multi-source data of the target area, a multi-source ecological feature vector is constructed on the ecological unit.

[0034] In one example of this disclosure, the construction of a multi-source ecological feature vector on the ecological unit based on multi-source data of the target area includes: constructing surface ecological parameters; the surface ecological parameters include one or more of vegetation and land parameters, water body and water quality parameters, soil and salinity parameters, morphology and topography parameters, and anthropogenic disturbance parameters; determining the parameter values ​​of the surface ecological parameters on the ecological unit based on the multi-source data; and constructing the multi-source ecological feature vector based on the parameter values.

[0035] Specifically, surface ecological parameters can be obtained in advance by feature extraction based on historical data, and then the values ​​corresponding to each surface ecological parameter can be determined based on multi-source data to classify each ecological unit. s Constructing multi-source ecological feature vectors ∈S X s .

[0036] Surface ecological parameters include vegetation and land parameters, such as NDVI (Normalized Difference Vegetation Index), EVI (Enhanced Vegetation Index), SAVI (Soil-Adjusted Vegetation Index), NDMI (Normalized Difference Moisture Index), and NBR (Normalized Burn Ratio); they also include water body and water quality parameters, such as NDWI (Normalized Difference Water Index) and MNDWI (Modified Normalized Difference Water Index). Index, including the improved normalized difference water index, reflectance ratios (blue, green, red, near-infrared), and texture; also includes soil and salinity parameters, such as S3 (salinity index); also includes morphological and topographic parameters, such as slope, curvature, and geomorphic position index (TPI); and also includes anthropogenic disturbance parameters, such as road and residential distance, and nighttime light intensity.

[0037] For parameters, annual or multi-year trend and anomalous features can be constructed in the time domain, such as STL decomposition or CUSUM and BFAST change point features. STL is an algorithm that decomposes time series data into three core components: trend, seasonal, and residual. CUSUM is a sequential analysis technique used to monitor small, persistent shifts in process mean, and it is particularly good at detecting small abrupt changes that may be missed by traditional control charts. BFAST is a change point detection method specifically designed for analyzing time series with seasonal and trend components, combining STL-style decomposition with change point detection.

[0038] In step S103, initial ecological problem identification data are calculated for each ecological unit based on the multi-source ecological feature vector.

[0039] Figure 2 This illustration schematically depicts a flowchart of a method for calculating initial ecological problem identification data in an exemplary embodiment of this disclosure. Figure 2 As shown, the specific steps for calculating the initial ecological problem identification data include: Step S201: Design a joint loss function based on the data categories of the initial ecological problem identification data; Step S203: Train the initial deep learning network model according to the joint loss function to obtain the target deep learning network model; Step S205: Input the multi-source ecological feature vector into the target deep learning network model to obtain the ecological problem identification data of each data category output by the target deep learning network model.

[0040] Specifically, a multimodal, multi-task deep learning network model can be pre-built to process multimodal data and simultaneously execute multiple deep learning tasks to output ecological problem identification data of various data categories.

[0041] In one embodiment of this disclosure, both the initial ecological problem identification data and the target ecological problem identification data contain two categories of data: ecological problem probability and continuous ecological indicator data, only the corresponding discrete units are different.

[0042] Regarding how to calculate the initial ecological problem identification data, one approach is to use ecological semantic segmentation tasks to calculate different types of ecological problems. j In ecological units s probability of occurrence at [location] p j ( s This can be used to identify ecological problems such as bare / degraded land, encroached areas, runoff convergence / siltation zones, polluted water bodies, invasive alien patches, and mining disturbance areas. On the other hand, continuous ecological indicators can be calculated through continuous feature regression tasks. r In ecological units s Values ​​for each time period r ( s ), including continuous ecological indicators such as the degree of eutrophication of water bodies. Regional erosion intensity and vegetation cover The corresponding proxy value is predicted through the feature mapping module.

[0043] In step S201, the design of the joint loss function based on the data type of the initial ecological problem identification data includes: A first loss function is constructed based on the class imbalance and hard samples of the semantic segmentation task, and the semantic segmentation task is used to calculate the probability of the ecological problem. The mean square error of the continuous ecological index data is used as the second loss function; The regularization loss term is used as the third loss function; the regularization loss term is either a consistent regularization loss term or a domain adaptive regularization loss term. The joint loss function is constructed based on the first loss function, the second loss function, and the third loss function.

[0044] Specifically, to balance training stability and performance consistency for both classification and regression tasks, a joint loss function is adopted:

[0045] Among them, the first item For the first loss function: In the formula, The weight coefficients for the j-th segmentation task; Let be the Dice loss for the j-th segmentation task, used to measure the overlap between the predicted segmentation mask and the real mask, and to address the problem of "class imbalance (such as small objects, few sample classes)"; Focal loss for the j-th segmentation task is used to address the training imbalance caused by "difficult samples (high proportion of easy-to-classify samples)" by reducing the loss weight of easy-to-classify samples and focusing on difficult-to-classify samples.

[0046] Second item For the second loss function: In the formula, Let be the weighting coefficient of the q-th continuous ecological indicator; For predicting indicator values, For reference index values, Let be the mean squared error (MSE) of the q-th continuous ecological indicator, used to enforce consistency in the feature space.

[0047] Third item The third loss function: In the formula, =Regularization weight coefficient, used to balance the contributions of "classification + regression loss" and "consistency / domain adaptation loss", to prevent overfitting to a single task and improve the alignment ability of multimodal feature space; L is regularization loss term, which can improve the alignment ability of multimodal feature space. If it is "consistency loss", it forces the feature representation of different modalities or branches to be consistent. If it is domain adaptation loss, it is used to reduce the feature distribution difference between the source domain and the target domain.

[0048] In step S203, an initial deep learning network model is constructed using Transformer-Unet as the backbone. The initial deep learning network model is trained using the dataset. Ecological element identification and representation are achieved through attention and feature sharing mechanisms. The loss function value of the joint loss function is calculated, and the model parameters in the initial deep learning network model are adjusted using the loss function value. Finally, the target deep learning network model is obtained after training.

[0049] In step S205, after constructing multi-source ecological feature vectors for ecological units in the target region, the multi-source ecological feature vectors are input into the target deep learning network model. The model then completes the ecological semantic segmentation task and the continuous feature regression task, ultimately obtaining the probabilities of ecological problems. p j ( s ), and the values ​​of various continuous ecological indicators. r ( s ).

[0050] In step S105, the target area is discretized into multiple engineering units according to the constructable objects of ecological restoration technology, and a bipartite map of the target area is constructed based on the ecological units and the engineering units.

[0051] Specifically, engineering unit e is a suitable construction area unit that divides the target area from the perspective of ecological restoration technology construction based on the constructable objects, such as constructable plots, river sections, slopes, ponds, etc.

[0052] Figure 3 This schematically illustrates a flowchart of a bipartite graph construction method according to an exemplary embodiment of this disclosure. For example... Figure 3 As shown, the specific process of constructing a bipartite graph is as follows: Step S301: Construct an ecological unit set based on the ecological unit, and construct an engineering unit set based on the engineering unit; Step S303: When the ecological unit and the engineering unit have a spatial relationship, establish an edge between the ecological unit and the engineering unit, and calculate the edge weight of the edge; the spatial relationship includes one or more of intersection, containment, adjacency, and connectivity; Step S305: Construct the bipartite graph based on the ecological unit set, the engineering unit set, the edges, and the edge weights.

[0053] Specifically, an ecological unit set is constructed based on ecological unit s. And construct a set of engineering units based on engineering unit e. If ecological unit With engineering unit If any one or more of the conditions of intersection, containment, adjacency, and connectivity in spatial association are met, then an edge is established. , This yields set E. For each engineering unit... e The set of connected ecological units is denoted as N ( e )={ s |( s ,e )∈ E}

[0054] In order to characterize the strength of the association, edge weights are introduced. , ∈[0, 1]. In configuring edge weights , The calculation method varies depending on the type of edge.

[0055] For the edges corresponding to planar engineering units, the area coverage weighting method is used for calculation, i.e.

[0056] In the formula, This represents the area of ​​the intersection of s and e. This represents the area of ​​e itself.

[0057] For linear edges, such as river segments, a length-coverage weighting method is used for calculation, i.e.

[0058] In the formula, This represents the length of the intersection of s and e. This represents the length of e itself.

[0059] For some edges, an example decay weight should be configured, i.e.

[0060] In the formula, This represents the distance between s and e. This is the distance decay parameter. The closer s and e are, the greater the weight and the stronger the association; the farther apart they are, the smaller the weight and the weaker the association.

[0061] For edges at hydrological or topographical connections, the weight can be flexibly set based on confluence paths, slope aspect consistency, or river network connectivity. Confluence paths refer to the paths through which water flows converge; the weight of an edge may be related to whether it lies on a confluence path, as they play a crucial role in hydrological processes. Slope aspect refers to the orientation of the topographic surface; if edges e and s have the same slope aspect, it indicates similar topographic features and a potentially strong correlation, thus their weights may be larger. River network connectivity refers to the connection between different river segments within a river network; if edges e and s are connected within the river network, it indicates they are hydrologically or topographically interconnected, and their weights may be larger.

[0062] Therefore, the target area is discretized into a two-layer "ecological-engineering" unit, and the objectified geographic units are segmented at multiple scales to establish a bipartite map G=(Neco , N eng E, , ).

[0063] In step S107, based on the bipartite graph, the initial ecological problem identification data on each of the ecological units is converted into target ecological problem identification data on each of the engineering units.

[0064] Specifically, when carrying out ecological restoration in the target area, it is necessary to implement targeted ecological restoration technologies for different ecological problems. Therefore, in order to facilitate the intuitive display of ecological problem identification data for the implementation of the technology, the ecological problem identification data at the ecological unit level can be converted into ecological problem identification data at the engineering unit level.

[0065] In one embodiment of this disclosure, the step of converting initial ecological problem identification data on each of the ecological units into target ecological problem identification data on each of the engineering units based on the bipartite graph includes: Based on the bipartite graph, edges including engineering units are determined to extract the corresponding ecological units and edge weights. For a given engineering unit, the weighted data of ecological problem identification on the corresponding ecological unit are summed to obtain the target ecological problem identification data on the engineering unit; wherein, the weighted data of ecological problem identification is the product of the initial ecological problem identification data on the ecological unit and the edge weight; The engineering units are traversed to obtain target ecological problem identification data for each engineering unit.

[0066] Specifically, the output generated by the multi-task deep learning network model at the ecological unit scale is projected onto the engineering unit scale using a bipartite graph. The transformation process is as follows: ,

[0067] In the formula, N ( e (This refers to the engineering unit in the two-part diagram) e A collection of interconnected ecological units.

[0068] In one embodiment of this disclosure, in order to visually display the ecological problem identification data, an ecological problem probability distribution map and a continuous distribution map of ecological indicators can be drawn.

[0069] Based on the above method, the distribution of identified ecological problems can be transformed into the risk or demand intensity of constructible objects, which facilitates the subsequent calculation and processing of ecological technology restoration. For example, it can be used as input for technology applicability, cost and time, and optimization solutions, thereby assisting project managers in making technical restoration decisions and promoting the efficient implementation of ecological protection and restoration work.

[0070] Figure 4 This schematic diagram illustrates the composition of an ecological problem identification device in an exemplary embodiment of the present disclosure, such as... Figure 4 As shown, the ecological problem identification device 400 may include a vector module 401, a calculation module 402, a bipartite graph module 403, and a transformation module 404. Wherein: Vector unit 401 is used to discretize the target area into multiple ecological units according to the grid, and construct a multi-source ecological feature vector on the ecological units based on the multi-source data of the target area; The calculation unit 402 is used to calculate the initial ecological problem identification data on each of the ecological units based on the multi-source ecological feature vector; The bipartite graph unit 403 is used to discretize the target area into multiple engineering units according to the constructable objects, and to construct a bipartite graph of the target area based on the ecological units and the engineering units; The conversion unit 404 is used to convert the initial ecological problem identification data on each of the ecological units into target ecological problem identification data on each of the engineering units based on the bipartite graph.

[0071] According to an exemplary embodiment of this disclosure, the bipartite graph unit 403 is further configured to construct an ecological unit set based on the ecological unit, and to construct an engineering unit set based on the engineering unit; When the ecological unit and the engineering unit have a spatial relationship, an edge is established between the ecological unit and the engineering unit, and the edge weight is calculated; the spatial relationship includes one or more of intersection, inclusion, adjacency, and connectivity; the bipartite graph is constructed based on the ecological unit set, the engineering unit set, the edge, and the edge weight.

[0072] According to an exemplary embodiment of this disclosure, the conversion unit 404 is further configured to determine the edges including engineering units based on the bipartite graph, so as to extract the ecological units and edge weights corresponding to the edges; for an engineering unit, sum the weighted data of ecological problem identification on the corresponding ecological unit to obtain the target ecological problem identification data on the engineering unit; wherein, the weighted data of ecological problem identification is the product of the initial ecological problem identification data on the ecological unit and the edge weights; traverse the engineering units to obtain the target ecological problem identification data on each engineering unit.

[0073] According to an exemplary embodiment of this disclosure, the computing unit 402 is further configured to design a joint loss function based on the data categories of the initial ecological problem identification data; train an initial deep learning network model according to the joint loss function to obtain a target deep learning network model; and input the multi-source ecological feature vector into the target deep learning network model to obtain the ecological problem identification data of each data category output by the target deep learning network model.

[0074] According to an exemplary embodiment of this disclosure, the initial ecological problem identification data includes ecological problem probability and continuous ecological indicator data. The calculation unit 402 is further configured to construct a first loss function based on class imbalance and hard samples of the semantic segmentation task, wherein the semantic segmentation task is used to calculate the ecological problem probability; use the mean square error of the continuous ecological indicator data as a second loss function; use a regularization loss term as a third loss function; wherein the regularization loss term is a consistency regularization loss term or a domain adaptive regularization loss term; and construct the joint loss function based on the first loss function, the second loss function, and the third loss function.

[0075] According to an exemplary embodiment of the present disclosure, the vector unit 401 is further configured to acquire multi-source data of the target area, the multi-source data including one or more of optical image data, radar image data and terrain data.

[0076] According to an exemplary embodiment of this disclosure, the vector unit 401 is further configured to construct surface ecological parameters; the surface ecological parameters include one or more of vegetation and land parameters, water body and water quality parameters, soil and salinity parameters, morphology and topography parameters, and anthropogenic disturbance parameters; determine the parameter values ​​of the surface ecological parameters on the ecological unit based on multi-source data; and construct the multi-source ecological feature vector based on the parameter values.

[0077] The specific details of each module in the aforementioned ecological problem identification device 400 have been described in detail in the corresponding ecological problem identification method, so they will not be repeated here.

[0078] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0079] In exemplary embodiments of this disclosure, a storage medium capable of implementing the above-described methods is also provided. It may be a portable compact disc read-only memory (CD-ROM) and include program code, and can run on a terminal device, such as a mobile phone. However, the program product of this disclosure is not limited thereto. In this document, a readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0080] In an exemplary embodiment of this disclosure, an electronic device capable of implementing the above-described method is also provided. Figure 5 The schematic diagram illustrates the structure of a computer system of an electronic device according to an exemplary embodiment of the present disclosure.

[0081] It should be noted that, Figure 5 The computer system 500 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0082] like Figure 5 As shown, the computer system 500 includes a Central Processing Unit (CPU) 501, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 502 or programs loaded from storage section 508 into Random Access Memory (RAM) 503. The RAM 503 also stores various programs and data required for system operation. The CPU 501, ROM 502, and RAM 503 are interconnected via a bus 504. An Input / Output (I / O) interface 505 is also connected to the bus 504.

[0083] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. Removable media 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 510 as needed so that computer programs read from them can be installed into storage section 508 as needed.

[0084] In particular, according to embodiments of this disclosure, the processes described below with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by central processing unit (CPU) 501, it performs various functions defined in the system of this disclosure.

[0085] It should be noted that the computer-readable medium shown in the embodiments of this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0086] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0087] The units described in the embodiments of this disclosure can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the unit itself.

[0088] In another aspect, this disclosure also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to perform the methods described in the above embodiments.

[0089] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0090] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0091] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein.

[0092] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for identifying ecological problems, characterized in that, include: The target area is discretized into multiple ecological units by grid, and a multi-source ecological feature vector is constructed on the ecological units based on the multi-source data of the target area; Initial ecological problem identification data are calculated for each ecological unit based on the multi-source ecological feature vector; The target area is discretized into multiple engineering units according to the constructable objects of ecological restoration technology, and a bipartite map of the target area is constructed based on the ecological units and the engineering units; Based on the bipartite graph, the initial ecological problem identification data in each of the ecological units will be converted into target ecological problem identification data in each of the engineering units.

2. The method for identifying ecological problems according to claim 1, characterized in that, The construction of a bipartite map of the target region based on the ecological units and the engineering units includes: An ecological unit set is constructed based on the ecological units, and an engineering unit set is constructed based on the engineering units; When the ecological unit and the engineering unit have a spatial relationship, an edge is established between the ecological unit and the engineering unit, and the edge weight of the edge is calculated; the spatial relationship includes one or more of the following: intersection, containment, adjacency, and connectivity; The bipartite graph is constructed based on the set of ecological units, the set of engineering units, the edges, and the edge weights.

3. The method for identifying ecological problems according to claim 1 or 2, characterized in that, The process of converting initial ecological problem identification data in each of the ecological units into target ecological problem identification data in each of the engineering units based on the bipartite graph includes: Based on the bipartite graph, edges including engineering units are determined to extract the corresponding ecological units and edge weights. For a given engineering unit, the weighted data of ecological problem identification on the corresponding ecological unit are summed to obtain the target ecological problem identification data on the engineering unit; wherein, the weighted data of ecological problem identification is the product of the initial ecological problem identification data on the ecological unit and the edge weight; The engineering units are traversed to obtain target ecological problem identification data for each engineering unit.

4. The method for identifying ecological problems according to claim 1, characterized in that, The initial ecological problem identification data calculated based on the multi-source ecological feature vectors for each ecological unit includes: Design a joint loss function based on the data categories of the initial ecological problem identification data; The target deep learning network model is obtained by training the initial deep learning network model according to the joint loss function. The multi-source ecological feature vectors are input into the target deep learning network model to obtain the ecological problem identification data for each data category output by the target deep learning network model.

5. The method for identifying ecological problems according to claim 4, characterized in that, The initial ecological problem identification data includes ecological problem probabilities and continuous ecological indicator data. The design of a joint loss function based on the data type of the initial ecological problem identification data includes: A first loss function is constructed based on the class imbalance and hard samples of the semantic segmentation task, and the semantic segmentation task is used to calculate the probability of the ecological problem. The mean square error of the continuous ecological index data is used as the second loss function; The regularization loss term is used as the third loss function; the regularization loss term is either a consistent regularization loss term or a domain adaptive regularization loss term. The joint loss function is constructed based on the first loss function, the second loss function, and the third loss function.

6. The method for identifying ecological problems according to claim 1, characterized in that, The method further includes: acquiring multi-source data of the target area, wherein the multi-source data includes one or more of optical image data, radar image data, and terrain data.

7. The method for identifying ecological problems according to claim 6, characterized in that, The multi-source ecological feature vector constructed on the ecological unit based on multi-source data of the target region includes: Construct surface ecological parameters; the surface ecological parameters include one or more of the following: vegetation and land parameters, water body and water quality parameters, soil and salinity parameters, morphological and topographic parameters, and anthropogenic disturbance parameters; The parameter values ​​of the surface ecological parameters on the ecological unit are determined based on multi-source data; The multi-source ecological feature vector is constructed based on the parameter values.

8. An ecological problem identification device, characterized in that, include: Vector unit, used to discretize the target area into multiple ecological units according to the grid, and to construct a multi-source ecological feature vector on the ecological unit based on the multi-source data of the target area; A computing unit is used to calculate initial ecological problem identification data on each of the ecological units based on the multi-source ecological feature vectors. The bipartite graph unit is used to discretize the target area into multiple engineering units according to the constructable objects of ecological restoration technology, and to construct a bipartite graph of the target area based on the ecological units and the engineering units; The conversion unit is used to convert the initial ecological problem identification data in each of the ecological units into target ecological problem identification data in each of the engineering units based on the bipartite graph.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the ecological problem identification method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the ecological problem identification method as described in any one of claims 1 to 7.

Images