A global space planting data management method and system

Abstract

This invention discloses a method and system for managing spatial planting data across the entire domain, belonging to the field of agricultural information technology. The method includes: establishing a multi-dimensional unified spatial coordinate system covering horizontal, underground, and vertical spaces; binding planting units to spatial coordinates to generate dynamic spatial labels; collecting spatial status data through distributed spatial sensing devices, dynamically updating labels, and recording spatial location change trajectories; standardizing spatial planting data for different agricultural scenarios; constructing a spatiotemporal map and outputting spatial analysis results. This invention achieves unified management of multi-dimensional spatial data and dynamic tracking of planting unit spatial trajectories through a multi-dimensional unified spatial coordinate system and a dynamic spatial labeling mechanism; breaks down data barriers between different agricultural scenarios through cross-scenario data standardization processing; and achieves joint analysis of time and spatial dimensions through spatiotemporal map construction, improving the intelligence level of planting management decisions.

Description

Technical Field

[0001] This application belongs to the field of agricultural information technology, specifically involving a method and system for managing planting data across the entire spatial domain. It is particularly suitable for scenarios that integrate agricultural planting data across multiple spatial dimensions, including ground horizontal space, underground space, and vertical space, for unified management. Background Technology

[0002] In recent years, with the deep application of technologies such as the Internet of Things, big data, and artificial intelligence in the agricultural field, smart agriculture has become an important direction for the modernization of agriculture. As an important component of smart agriculture, planting data management undertakes core functions such as environmental monitoring, crop growth tracking, and resource allocation. Its data management capabilities directly affect the effectiveness of precision planting and intelligent decision-making.

[0003] In actual agricultural production, planting scenarios are becoming increasingly diversified. In addition to traditional horizontal planting, new planting models such as vertical agriculture, three-dimensional cultivation, and rooftop planting are developing rapidly, and planting spaces are showing a trend towards being three-dimensional and complex. At the same time, the data management needs of planting in different scenarios such as greenhouse planting, field planting, and facility agriculture also differ, posing challenges to the unified management of planting data.

[0004] Existing agricultural data management systems are widely used in the areas of planting data collection, storage, and analysis. For example, greenhouse environmental monitoring systems can achieve real-time collection and early warning of environmental parameters within greenhouses; field management systems can utilize positioning technology to achieve spatial monitoring of farmland; and vertical farming systems can realize the logistics scheduling of crops in vertical space. These systems have played a certain role in their respective application scenarios.

[0005] However, the inventors discovered that existing technologies still have some shortcomings in planting data management. On the one hand, existing systems are mostly designed for a single scenario or a single spatial dimension, resulting in fragmented data management when planting involves multiple spatial forms or agricultural scenarios. On the other hand, existing technologies have limited ability to track dynamic changes in crop spatial location, making it difficult to support spatial trajectory analysis throughout the entire crop life cycle. Furthermore, data between different systems is difficult to exchange, affecting cross-scenario and cross-spatial data analysis and decision-making.

[0006] Therefore, it is necessary to study an improved method and system for managing planting data in order to enhance the efficiency of planting data management and analysis capabilities. Summary of the Invention

[0007] The purpose of this application is to provide a method and system for managing planting data across the entire domain, in order to solve the problems of fragmented data management, limited spatial location tracking capabilities, and difficulties in cross-scenario data interoperability in the existing technology.

[0008] To achieve the above objectives, the first aspect of this application provides a method for managing spatial planting data across the entire domain, comprising the following steps: Establish a multi-dimensional unified spatial coordinate system for the entire planting area. The multi-dimensional unified spatial coordinate system covers the horizontal space, underground space and vertical space of the entire planting area, and assigns a unique multi-dimensional spatial coordinate identifier to each spatial location within the entire planting area. The planting unit is bound to the multidimensional spatial coordinate identifier corresponding to the planting unit in the multidimensional spatial unified coordinate system, and a dynamic spatial label including the binding relationship between the planting unit and its corresponding multidimensional spatial coordinate identifier is generated. Spatial status data of each planting unit is collected by distributed spatial sensing devices. The multi-dimensional spatial coordinate identifier in the dynamic spatial label of the planting unit is updated according to the spatial status data, and the spatial position change trajectory of the planting unit is recorded. Acquire spatial planting data under different agricultural scenarios, and standardize the spatial planting data based on the multi-dimensional unified spatial coordinate system to obtain standardized spatial planting data. A global spatial planting database is constructed based on the dynamic spatial labels, the spatial state data, the spatial location change trajectory, and the standardized spatial planting data. A spatiotemporal map is constructed based on the global spatial planting database, and spatial analysis results of planting units are output based on the spatiotemporal map.

[0009] By establishing a unified multi-dimensional spatial coordinate system covering horizontal, underground, and vertical spaces, and binding planting units with spatial coordinates to generate dynamic spatial labels, planting data from different spatial dimensions can be uniformly mapped to the same coordinate system, eliminating spatial data silos and achieving unified management of multi-dimensional spatial data. By recording the spatial location change trajectory of planting units, the spatial migration process of planting units throughout their entire life cycle can be traced, supporting the analysis of the planting process. By constructing a spatiotemporal map and outputting spatial analysis results, data from the time and spatial dimensions can be jointly analyzed, improving the level of intelligence in planting decision-making.

[0010] Preferably, establishing a unified coordinate system in multidimensional space includes: Multidimensional spatial coordinates are represented by a quadruple: (X, Y, Z, T), where X and Y represent the horizontal coordinates on the ground, Z represents the vertical coordinates, and the positive and negative values ​​of Z represent the vertical height and underground depth, respectively, and T represents the time coordinate. The entire planting area is divided into multiple spatial grid units according to a preset spatial resolution, and a unique spatial grid code is assigned to each spatial grid unit.

[0011] By adopting a quadruples coordinate representation and spatial grid partitioning, any spatial location can be represented and indexed in a unified numerical form, improving the accuracy of spatial positioning and the efficiency of spatial query.

[0012] Preferably, the step of binding the planting unit with its corresponding multidimensional spatial coordinate identifier in a unified multidimensional spatial coordinate system to generate a dynamic spatial label including the binding relationship between the planting unit and its corresponding multidimensional spatial coordinate identifier includes: Assign a unique identifier to the planting unit and obtain the initial multidimensional spatial coordinate identifier of the planting unit; The unique identifier of the planting unit is bound to the initial multidimensional spatial coordinate identifier to generate an initial dynamic spatial label and store it.

[0013] By assigning a unique identifier to each planting unit and binding it to spatial coordinates, each planting unit acquires a unique spatial identity, laying the foundation for subsequent spatial tracking and source analysis.

[0014] Preferably, the step of collecting spatial status data of each planting unit through distributed spatial sensing devices includes: The horizontal coordinates of the planting unit are collected using a ground position sensor; The vertical height or underground depth coordinates of the planting unit are collected by a vertical position sensor; Temperature, humidity, light, or nutrient concentration parameters of the space where the planting unit is located are collected by spatial environment sensors; The spatial adjacency relationship of the planting units is determined based on the spatial distance between each planting unit.

[0015] By using multiple types of sensors to collect data collaboratively, the spatial status data of the planting unit covers multiple dimensions such as location, environment, and relationships, thus improving the comprehensiveness and accuracy of spatial perception.

[0016] Preferably, updating the multi-dimensional spatial coordinate identifier in the dynamic spatial label of the planting unit based on spatial state data, and recording the spatial position change trajectory of the planting unit, includes: Compare the current multidimensional spatial coordinate identifier of the planting unit with the historical coordinate identifier, and calculate the spatial position change. The historical coordinate identifier includes the initial multidimensional spatial coordinate identifier of the planting unit. When the change in spatial location exceeds a preset threshold, the multidimensional spatial coordinate identifier of the planting unit is updated and a migration record is added to the spatial trajectory table. When the change does not exceed the threshold, the environmental parameters in the dynamic spatial label are updated according to the changes in environmental parameters.

[0017] By comparing spatial location changes in real time and dynamically updating the labels, the dynamic spatial labels can reflect the true spatial status of the planting units, ensuring the timeliness and accuracy of the spatial data.

[0018] Preferably, the step of acquiring spatial planting data under different agricultural scenarios, and standardizing the spatial planting data based on the multi-dimensional spatial unified coordinate system to obtain standardized spatial planting data, includes: Acquire spatial planting data, and identify agricultural scene types based on the spatial planting data. The agricultural scene types include greenhouse scenes, field scenes, vertical agriculture scenes, or rooftop planting scenes. Based on the agricultural scenario type, obtain the corresponding coordinate system transformation rules and convert the local coordinate system of the spatial planting data into the multidimensional spatial coordinate values ​​of the multidimensional unified spatial coordinate system. The converted spatial planting data is subjected to spatial consistency verification, and the spatial planting data that passes the verification is stored in the global spatial planting database.

[0019] By using scene recognition and coordinate system transformation, spatial data from different agricultural scenarios can be uniformly mapped to a multi-dimensional unified coordinate system, eliminating obstacles to cross-scenario data fusion and achieving unified management of spatial data across scenarios.

[0020] Preferably, the construction of the spatiotemporal map based on the global spatial planting database includes: Calculate the spatial distance between each planting unit. When the spatial distance is less than a preset adjacency distance threshold, it is determined that two planting units have a spatial adjacency relationship. A spatial adjacency graph structure is constructed using the planting units as nodes and the spatial adjacency relationships between the planting units as edges; Organize the spatial status data change sequence of each planting unit according to time sequence to construct time series data; The spatial adjacency graph structure and the time series data are input into the spatiotemporal graph neural network, and the spatiotemporal graph node representation is output.

[0021] By constructing a spatial adjacency graph structure and time series data and inputting them into a spatiotemporal graph neural network, spatial and temporal relationships can be integrated and modeled, providing a data foundation for spatiotemporal joint analysis.

[0022] Preferably, the spatiotemporal graph neural network includes spatial convolutional layers, temporal convolutional layers, and spatiotemporal attention layers: The spatial convolutional layer is used to perform graph convolution operations on the spatial adjacency graph structure to extract the spatial adjacency features of the planting units; The temporal convolutional layer is used to perform temporal convolution operations on the time series data to extract the temporal variation features of the planting units. The spatiotemporal attention layer is used to fuse the spatial adjacency features and the temporal variation features, and output a spatiotemporal graph node representation that fuses spatial and temporal relationships.

[0023] By working together with spatial convolutional layers, temporal convolutional layers, and spatiotemporal attention layers, the effectiveness of spatiotemporal feature extraction is improved, thereby enhancing the accuracy of the spatiotemporal prediction model.

[0024] Preferably, the step of outputting the spatial analysis results of the planting unit based on the spatiotemporal map includes: Based on the aforementioned spatiotemporal map, the transmission path of pests and diseases can be traced according to the temporal sequence and spatial location of their occurrence. Based on the spatiotemporal map, the spatial location of the planting unit at future time points is predicted according to the historical spatial trajectory sequence and current spatial state data of the planting unit, and the growth trajectory prediction result is generated.

[0025] By using spatiotemporal maps to trace the spread of pests and diseases and predict their growth trajectories, intelligent support is provided for planting management decisions, which helps with pest and disease control and harvest planning.

[0026] A second aspect of this application also provides a comprehensive spatial planting data management system, comprising: The multi-dimensional space coordinate management module is used to establish and manage the aforementioned unified multi-dimensional space coordinate system. The planting unit spatial binding module is used to bind planting units to multi-dimensional spatial coordinate identifiers and generate the dynamic spatial labels; A distributed spatial sensing layer is used to collect spatial status data of each planting unit; The dynamic spatial label management module is used to update the multi-dimensional spatial coordinate identifiers in the dynamic spatial labels and record the trajectory of spatial location changes. The cross-scenario data standardization module is used to standardize spatial planting data in different agricultural scenarios; The global spatial planting database construction module is used to construct the global spatial planting database; The spatiotemporal map construction module is used to construct a spatiotemporal map based on the global spatial planting database; The spatial analysis module is used to output the spatial analysis results of the planting units based on the spatiotemporal map; The system is configured to perform the method described in any of the first aspects above.

[0027] Compared with the prior art, this application has the following advantages: (1) By establishing a unified coordinate system for multidimensional space, unified management of multidimensional spatial data in ground horizontal space, underground space and vertical space has been realized, eliminating spatial data silos; (2) Through the dynamic spatial tagging mechanism, the dynamic tracking of the spatial location of planting units is realized, and the spatial trajectory tracing of the planting units throughout their entire life cycle is supported; (3) Through cross-scenario data standardization processing, spatial data fusion of different agricultural scenarios such as greenhouses, open fields, and vertical agriculture has been achieved, breaking down data barriers between scenarios; (4) By constructing a spatiotemporal graph and fusing spatiotemporal features based on a spatiotemporal graph neural network, joint analysis of time and space dimensions was achieved, which improved the intelligence level of planting decision-making. Attached Figure Description

[0028] The specific embodiments of this application will be further described in detail below with reference to the accompanying drawings. It should be understood that the described drawings are only for explaining this application and do not constitute any undue limitation on the scope of protection of this application.

[0029] Figure 1 This is a flowchart of the whole-domain spatial planting data management method provided in the embodiments of the present invention; Figure 2 This is a structural block diagram of the whole-domain spatial planting data management system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the spatiotemporal map construction process provided in the embodiments of the present invention. Detailed Implementation

[0030] The specific embodiments of this application will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustrative purposes only and do not constitute any undue limitation on the scope of protection of this application. Those skilled in the art, after reading the above-described summary of this application, will be fully capable of implementing the technical solutions of this application.

[0031] In the following description, for clarity and brevity, detailed descriptions of known functions and constructions have been omitted. The same reference numerals in the accompanying drawings consistently denote the same elements.

[0032] See Figure 1 This is a flowchart of the whole-domain spatial planting data management method provided in the embodiments of this application. Figure 1 As shown, the whole-domain spatial planting data management method provided by the present invention includes the following steps: In this step, it is necessary to establish a unified multi-dimensional spatial coordinate system covering the horizontal space, underground space and vertical space, and assign a unique multi-dimensional spatial coordinate identifier to each spatial location within the entire planting area.

[0033] Specifically, the multidimensional spatial unified coordinate system uses a quadruple form to represent multidimensional spatial coordinates: (X, Y, Z, T). Here, X and Y represent the horizontal coordinates of the ground, used to characterize the spatial distribution of the planting area on the horizontal plane; Z represents the vertical coordinate, with positive Z representing vertical height and negative Z representing underground depth, used to characterize the vertical height of vertical cultivation racks and greenhouses, or the vertical depth of soil profiles and root layers; T represents the time coordinate, used to characterize the temporal position of the planting unit within its growth cycle.

[0034] When establishing a coordinate system, the origin and axis directions must first be determined. The origin can be the geometric center of the entire planting area, the entrance point, or another easily manageable location. The X-axis can point due east, the Y-axis can point due north, and the Z-axis is positive upwards and negative downwards. The coordinate unit can be determined according to actual needs; for example, meters can be used as the unit of length, and the ISO 8601 standard time format can be used as the unit of time.

[0035] After determining the coordinate system, it is also necessary to determine the spatial resolution based on the accuracy requirements of planting management. The entire planting area is then divided into multiple spatial grid units according to the spatial resolution, and a unique spatial grid code is assigned to each spatial grid unit. Spatial grid coding can adopt Morton coding, Hilbert curve coding, or simple row and column number coding, etc., to support efficient spatial query operations.

[0036] For example, in an embodiment of an agricultural technology park, the park simultaneously operates greenhouse cultivation areas and vertical farming areas. The park defines a unified multi-dimensional spatial coordinate system as follows: the ground horizontal coordinates (X, Y) adopt the park's local coordinate system, with the origin at the park entrance, the X-axis pointing due east, and the Y-axis pointing due north, in meters; in the vertical coordinates (Z), Z=0 represents the ground, Z>0 represents vertical height, and Z<0 represents underground depth; the time coordinates (T) adopt the ISO 8601 standard time format. The coordinate range is set as follows: X is -50 meters to +50 meters, Y is -50 meters to +50 meters, and Z is -2 meters to +10 meters (2 meters underground to 10 meters above ground). The spatial resolution is set to 1 meter, meaning each spatial grid unit is a 1-meter × 1-meter × 1-meter cube.

[0037] Step S102: Bind the planting unit to the multi-dimensional spatial coordinate identifier to generate a dynamic spatial label.

[0038] In this step, the planting units to be managed need to be spatially bound to a unified multi-dimensional coordinate system, and a dynamic spatial label corresponding to the spatial location of each planting unit needs to be generated.

[0039] A planting unit refers to the smallest unit for planting management, which can be a single plant, a planting container (such as a flowerpot or planting bag), a planting area (such as a plot or field), or other forms of planting object. The granularity of the planting unit can be flexibly adjusted according to actual management needs. For example, in a greenhouse with intensive management, each plant can be considered a planting unit; in a large-scale field, each plot or each management unit can be considered a planting unit.

[0040] Dynamic spatial tags are one of the core concepts of this invention, used to establish a dynamic binding relationship between planting units and spatial coordinates. A dynamic spatial tag includes the binding relationship between a planting unit identifier and a multi-dimensional spatial coordinate identifier, and may also include auxiliary information such as tag creation time, tag update time, tag validity period, and spatial status markers.

[0041] In practice, each planting unit is first assigned a unique planting unit identifier (ID). The planting unit identifier can be generated using coding rules, such as the format "scene code-area code-serial number," where "WT-001" represents planting unit number 1 in the greenhouse scene, and "VA-001" represents planting unit number 1 in the vertical agriculture scene. The planting unit identifier must be unique throughout the entire planting area to support accurate identification and tracking of planting units.

[0042] Next, the initial multidimensional spatial coordinates of the planting unit are obtained. These initial coordinates can be obtained manually, collected by positioning equipment, or calculated based on the initial position of the planting unit. For example, in a grid divided on the greenhouse floor, the coordinates of the center point of each grid cell can be used as the initial coordinates of the planting unit within that cell; in a vertical farming system, the position of the planting trough can be calculated based on the position and height of the cultivation rack.

[0043] Finally, the planting unit identifier is bound to the initial multidimensional spatial coordinate identifier to generate and store the initial dynamic spatial label. The dynamic spatial label can be stored in a planting unit spatial mapping table, which records the correspondence between the planting unit identifier and the current multidimensional spatial coordinate identifier, supporting fast query and update operations.

[0044] Preferably, the data structure of the dynamic spatial tag includes, but is not limited to, the following fields: a unique identifier (ID) for the planting unit, current multidimensional spatial coordinates (X, Y, Z, T), tag creation timestamp, most recent update timestamp, tag validity period marker, and current spatial status marker. This data structure is stored in the form of database records or data objects, providing a unified data carrier for spatial tracking and analysis.

[0045] Step S103: Collect spatial status data of each planting unit through distributed spatial sensing devices.

[0046] In this step, distributed spatial sensing devices are deployed to collect real-time spatial status data for each planting unit. This spatial status data includes spatial location data, spatial relationship data, and spatial environment data.

[0047] Spatial location data refers to the current coordinate position of the planting unit in a multi-dimensional unified coordinate system, including horizontal coordinates (X, Y) and vertical height or depth coordinates (Z). Spatial location data can be acquired using ground-based and vertical position sensors. Ground-based sensors, including GNSS positioning modules, RFID readers, and Ultra Wide Band (UWB) positioning base stations, can acquire the horizontal coordinates of the planting unit. Vertical position sensors, including laser rangefinders, ultrasonic rangefinders, and encoders, can acquire the vertical height or depth coordinates of the planting unit. For planting units in fixed locations (such as planting troughs fixed on cultivation racks), their spatial position can be determined based on initially set position parameters, eliminating the need for real-time data acquisition.

[0048] Spatial relationship data refers to the spatial adjacency and distance relationships between planting units and other planting units. Spatial adjacency can be determined based on the spatial distance between planting units; when the spatial distance between two planting units is less than a preset adjacency distance threshold, the two planting units are considered to be spatially adjacent. The adjacency distance threshold can be set according to planting management needs, such as setting it as the standard spacing between adjacent planting units. Spatial relationship data is of significant value for applications such as pest and disease transmission analysis and resource scheduling optimization.

[0049] Spatial environmental data refers to environmental parameters such as temperature, humidity, light intensity, and nutrient concentration at the location of the planting unit. This data can be collected using spatial environmental sensors deployed at various locations, including temperature sensors, humidity sensors, light intensity sensors, and nutrient concentration sensors. The deployment density of these sensors can be determined based on monitoring accuracy requirements and cost budgets; for example, one sensor node can be deployed every 25 square meters in a greenhouse environment.

[0050] Distributed spatial sensing devices are deployed using a distributed architecture. Each sensor node connects to the data center via wired or wireless networks, enabling real-time data transmission and centralized processing. Sensor nodes can possess edge computing capabilities, performing data preprocessing locally (such as data cleaning and outlier filtering) before uploading, reducing the burden on the data center and improving system response speed.

[0051] Step S104: Update the dynamic spatial label based on the spatial status data and record the trajectory of spatial location changes.

[0052] In this step, the dynamic spatial labels of each planting unit are dynamically updated based on the real-time collected spatial status data, and the spatial location change trajectory of the planting unit is recorded.

[0053] A dynamic update mechanism is crucial for ensuring the timeliness of spatial data. In practice, it requires comparing the current multidimensional spatial coordinates of planting units with their historical coordinates in real-time or periodically to calculate changes in spatial location. Historical coordinates include the initial multidimensional spatial coordinates of the planting unit and the coordinates recorded during the most recent update.

[0054] When the change in spatial location exceeds a preset threshold, a spatial migration event is determined to have occurred in the planting unit. A migration record needs to be added to the spatial trajectory table, and the multidimensional spatial coordinates in the dynamic spatial label needs to be updated. The spatial trajectory table records the history of spatial location changes of the planting unit throughout its entire growth cycle, including migration time, original multidimensional spatial coordinates, and target multidimensional spatial coordinates. The spatial trajectory table provides data support for subsequent spatial trajectory queries and growth process analysis.

[0055] The spatial trajectory table is a database table used to persistently store the history of spatial location changes of planting units, and its records are associated with the dynamic spatial tags. When a spatial migration event is determined, the migration record appended to the spatial trajectory table includes at least: migration event time, planting unit identifier, coordinate identifier before migration (i.e., historical coordinate identifier), and coordinate identifier after migration (i.e., current coordinate identifier). The spatial trajectory table is associated with the dynamic spatial tag table through the planting unit identifier, thereby enabling querying from the current state to the complete historical trajectory.

[0056] When the change in spatial location does not exceed the threshold, it indicates that the planting unit has not migrated spatially, but environmental parameters may have changed. In this case, the environmental parameters in the dynamic spatial label can be updated based on the changes in environmental parameters. The environmental parameters can be updated incrementally or fully. Incremental updates only record the changed environmental parameters, while full updates replace all environmental parameters.

[0057] In addition, to prevent dynamic spatial tags from becoming invalid due to prolonged lack of updates, an expiration date can be set for them. The expiration date includes the tag creation time, tag update time, and valid status. When the current time exceeds the expiration period (e.g., more than 24 hours without updating), the dynamic spatial tag is marked as invalid, and a re-collection command is sent to the distributed spatial sensing device, triggering the re-collection of spatial status data. The expiration date should be determined based on the migration characteristics of the planting unit and the required data update frequency. For planting units that migrate frequently, the expiration date should be shorter; for planting units with fixed locations, the expiration date can be longer.

[0058] Step S105: Obtain spatial planting data under different agricultural scenarios and perform standardization processing.

[0059] In this step, spatial planting data from different agricultural scenarios such as greenhouses, open fields, vertical agriculture, and rooftop planting are acquired. The spatial planting data is then standardized based on a multi-dimensional spatial unified coordinate system to obtain standardized spatial planting data.

[0060] Spatial planting data varies significantly across different agricultural scenarios. In greenhouse scenarios, spatial planting data typically uses the greenhouse building structure as the coordinate system reference, and the data format may be specific to greenhouse management systems. In open-field scenarios, spatial planting data usually uses plot boundaries as the coordinate system reference, and the data format may be a GIS standard format. In vertical agriculture scenarios, spatial planting data typically uses cultivation rack levels as the spatial index, and the data format may be specific to facility agriculture systems. In rooftop planting scenarios, spatial planting data typically uses the roof plane as the coordinate system reference. These differences make cross-scenario data fusion difficult and require standardization to eliminate them.

[0061] Standardization processing includes steps such as scene recognition, coordinate system transformation, data format standardization, and spatial consistency verification.

[0062] Scene recognition refers to identifying agricultural scene types based on the source identifier, data format characteristics, or spatial coordinate range of spatial planting data. Scene recognition can be achieved using methods such as rule matching and machine learning. For example, the scene type can be directly identified based on the scene code in the data source identifier, or the scene type can be inferred based on data format characteristics (such as field names and data structures), or the scene type can be determined based on the spatial coordinate range (such as whether the coordinate values ​​include vertical height).

[0063] Coordinate system transformation refers to obtaining the corresponding coordinate system transformation rules based on the agricultural scenario type, and converting the local coordinate system of spatial planting data into multidimensional spatial coordinate values ​​of a unified multidimensional spatial coordinate system. The coordinate system transformation rules for different scenarios need to be predefined and configured. For example, the local coordinate system of a greenhouse scenario is based on the greenhouse building structure; when converting to a unified coordinate system, the actual geographical location of the greenhouse in the global space needs to be considered. The local coordinate system of a field scenario is based on the plot boundary; when converting to a unified coordinate system, the spatial layout of the plot boundary and planting rows / columns within the plot needs to be considered. The local coordinate system of a vertical agriculture scenario is based on the cultivation rack level; when converting to a unified coordinate system, the cultivation rack level height needs to be extracted as the vertical coordinate. The local coordinate system of a rooftop planting scenario is based on the roof plane; when converting to a unified coordinate system, the roof elevation needs to be considered as the reference height.

[0064] Data format standardization refers to converting spatial data formats from different scenarios into a unified spatial data format. A unified spatial data format should include information such as basic attributes of planting units (e.g., identifier, variety, planting date, growth stage), spatial location attributes (e.g., unified coordinates in multi-dimensional space), spatial relationship attributes (e.g., the scene to which it belongs, a list of adjacent spatial units), and spatial environment attributes (e.g., a list of environmental parameters and sampling timestamps).

[0065] Spatial consistency verification refers to validating the transformed spatial data to detect issues such as spatial overlap or spatial conflicts. Spatial overlap occurs when two or more planting units occupy the same spatial location, which may be due to data errors or transformation errors. Spatial conflicts occur when the multidimensional spatial coordinates of a planting unit exceed the boundaries of the valid planting area, which may be due to coordinate system transformation errors. Spatial data that fails verification should be marked as abnormal data and subject to manual review or automatic correction.

[0066] Constructing a spatiotemporal graph first requires building a spatial adjacency graph structure. Using planting units as nodes and the spatial adjacency relationships between planting units as edges, a spatial adjacency graph structure is constructed. The method for determining spatial adjacency relationships is as follows: calculate the spatial distance between each planting unit; when the spatial distance is less than a preset adjacency distance threshold, it is determined that two planting units have a spatial adjacency relationship. The adjacency distance threshold can be set according to the type of planting unit and planting management needs. For example, for adjacent planted plants, the adjacency distance threshold can be set to 1.5 times the plant crown width; for planting containers, the adjacency distance threshold can be set to the container spacing plus a certain margin.

[0067] Next, time-series data is constructed. Using the growth cycle of the planting unit as the time axis, the spatial state data of the planting unit is organized chronologically. The time-series data records information such as the spatial location, environmental parameters, and growth status of the planting unit at different time points, reflecting the spatiotemporal evolution of the planting unit.

[0068] Finally, the spatial adjacency graph structure and time series data are input into the spatiotemporal graph neural network, which outputs a spatiotemporal graph node representation. The spatiotemporal graph neural network is a neural network model capable of simultaneously processing graph structure data and time series data. The spatiotemporal graph neural network used in this invention includes spatial convolutional layers, temporal convolutional layers, and spatiotemporal attention layers.

[0069] Spatial convolutional layers perform graph convolution operations on the spatial adjacency graph structure to extract the spatial adjacency features of planting units. Graph convolution operations can aggregate feature information from neighboring nodes and capture feature patterns in spatial adjacency relationships. Temporal convolutional layers perform temporal convolution operations on time-series data to extract the temporal variation features of planting units. Temporal convolution operations can capture local patterns and long-term dependencies in time series data. Spatiotemporal attention layers fuse spatial adjacency features and temporal variation features, outputting a spatiotemporal graph node representation that integrates spatial and temporal relationships. The attention mechanism can adaptively assign importance weights to different features, improving the effectiveness of feature fusion.

[0070] Based on spatiotemporal graph node representations, various spatial analysis applications can be performed. For example, the spread of pests and diseases can be traced according to their temporal sequence and spatial location. Pests and diseases typically start from an infection source and gradually spread to adjacent planting units according to spatial adjacency. Spatiotemporal graphs can record this spread process, supporting pest and disease control decisions. Another example is predicting the spatial location of planting units at future points in time based on their historical spatial trajectory sequences and current spatial status data, generating growth trajectory prediction results. Growth trajectory prediction can help planting managers plan harvest time in advance, schedule harvesting equipment, and arrange logistics and transportation.

[0071] Figure 3 This is a schematic diagram of the spatiotemporal map construction process provided in an embodiment of the present invention. For example... Figure 3 As shown, the spatiotemporal graph construction process includes three stages: spatial adjacency graph construction, time series construction, and spatiotemporal fusion. In the spatial adjacency graph construction stage, spatial adjacency relationships are calculated based on the spatial location of planting units, constructing a graph structure with planting units as nodes and spatial adjacency relationships as edges. In the time series construction stage, the spatial state data of each planting unit is organized in chronological order to form time series data. In the spatiotemporal fusion stage, the spatial adjacency graph and time series data are input into a spatiotemporal graph neural network, and after spatial convolution, temporal convolution, and spatiotemporal attention fusion, the spatiotemporal graph node representation is output.

[0072] See Figure 2 ,like Figure 2The whole-domain spatial planting data management system provided in this application includes: a multi-dimensional spatial coordinate management module, a planting unit spatial binding module, a distributed spatial perception layer, a dynamic spatial label management module, a cross-scene data standardization module, a whole-domain spatial planting database construction module, a spatiotemporal map construction module, and a spatial analysis module.

[0073] The multidimensional spatial coordinate management module is used to establish and manage a unified multidimensional spatial coordinate system. Its main functions include: defining the coordinate axes and coordinate ranges of the unified multidimensional spatial coordinate system; establishing coordinate mapping rules to convert spatial coordinates from different sources into unified multidimensional spatial coordinates; verifying the validity and uniqueness of coordinates; and maintaining the metadata information of the coordinate system. The multidimensional spatial coordinate management module is the foundation of the entire system, providing a unified spatial coordinate reference framework for other modules.

[0074] The planting unit spatial binding module is used to bind planting units to multi-dimensional spatial coordinate identifiers, generating dynamic spatial labels. The main functions of this module include: assigning a unique identifier to each planting unit; obtaining the initial spatial coordinates of each planting unit; binding the planting unit identifier to the initial spatial coordinates to generate an initial dynamic spatial label; and storing the initial dynamic spatial label in the planting unit spatial mapping table. The planting unit spatial binding module establishes the association between the identity of a planting unit and its spatial location, laying the foundation for subsequent spatial tracking.

[0075] The distributed spatial sensing layer is used to collect spatial status data from each planting unit. This layer comprises multiple distributed spatial sensing devices, such as ground position sensors, vertical position sensors, and spatial environment sensors. Its main functions include: real-time acquisition of spatial location data, spatial relationship data, and spatial environment data of the planting units; preprocessing the collected data (e.g., data cleaning, outlier filtering); and uploading the processed data to the data center. The distributed spatial sensing layer is the system's data source, and its sensing capabilities directly impact the system's data quality and application effectiveness.

[0076] The dynamic spatial tag management module is used to update the multi-dimensional spatial coordinate identifiers in dynamic spatial tags and record the trajectory of spatial location changes. The main functions of this module include: receiving spatial status data uploaded by the distributed spatial perception layer; comparing current coordinates with historical coordinates to detect spatial location changes; updating dynamic spatial tags and appending trajectory records when spatial migration is detected; and managing the validity period and status of dynamic spatial tags. The dynamic spatial tag management module is one of the core modules of the system, enabling dynamic tracking of the spatial location of planting units.

[0077] The cross-scenario data standardization module is used to standardize spatial planting data from different agricultural scenarios. Its main functions include: identifying the scenario type of the input data; obtaining coordinate system transformation rules based on the scenario type and performing coordinate transformation; converting data formats from different scenarios to a unified format; performing spatial consistency verification on the transformed data; and storing the verified data into a global spatial planting database. The cross-scenario data standardization module achieves the fusion of cross-scenario data, breaking down data barriers between scenarios.

[0078] The global spatial planting database construction module is used to build the global spatial planting database. Its main functions include: designing the database schema and data model; storing dynamic spatial labels, spatial status data, spatial location change trajectories, and standardized spatial planting data; establishing data indexes to support efficient queries; and maintaining data integrity and consistency. The global spatial planting database serves as the system's data center, providing data support for upper-layer applications.

[0079] The spatiotemporal graph construction module is used to construct spatiotemporal graphs based on a global spatial planting database. The main functions of this module include: calculating the spatial adjacency relationships between planting units and constructing a spatial adjacency graph structure; organizing spatial state data in chronological order to construct time-series data; training a spatiotemporal graph neural network model; and generating spatiotemporal graph node representations. The spatiotemporal graph construction module achieves fusion modeling of spatiotemporal data, providing a data foundation for spatiotemporal joint analysis.

[0080] The spatial analysis module outputs spatial analysis results for planting units based on spatiotemporal maps. Its main functions include: providing a spatiotemporal map query interface, supporting spatial range queries, temporal range queries, and spatiotemporal joint queries; performing spatiotemporal correlation analysis based on the spatiotemporal map, such as tracing pest and disease transmission and predicting growth trajectories; and outputting analysis results, supporting visualization and report generation. The spatial analysis module is the system's application layer, transforming data into decision support information.

[0081] The modules mentioned above communicate with each other through data interfaces, forming a complete data processing pipeline. The multi-dimensional spatial coordinate management module provides coordinate references for other modules; the planting unit spatial binding module generates initial dynamic spatial labels; the distributed spatial perception layer collects spatial status data; the dynamic spatial label management module updates dynamic spatial labels and trajectory records; the cross-scene data standardization module integrates data from multiple scenes; the global spatial planting database construction module stores and manages data; the spatiotemporal map construction module constructs spatiotemporal maps; and the spatial analysis module outputs analysis results.

[0082] Specifically, in the embodiments of this application, the multi-dimensional spatial unified coordinate system covers three dimensions: ground horizontal space, underground space, and vertical space.

[0083] In the horizontal spatial dimension, the coordinate system uses (X, Y) two-dimensional coordinates to represent the spatial distribution of the planting area on the horizontal plane. The directions of the X and Y axes can be set according to actual needs, typically choosing due east as the positive direction of the X-axis and due north as the positive direction of the Y-axis. The origin of the coordinate system can be the geometric center of the planting area, the entrance point, or other easily manageable location. The unit of the horizontal coordinates is usually meters, and the precision can be set according to management requirements.

[0084] In the subsurface spatial dimension, the coordinate system uses the negative Z-axis region to represent the vertical depth of the soil profile or root layer. The origin of the Z-axis (Z=0) is usually set at ground level, and Z<0 indicates subsurface depth. For example, Z=-0.5 represents a depth of 0.5 meters, corresponding to the root distribution layer in the soil profile. Subsurface spatial coordinates are of great significance for applications such as root monitoring and soil analysis.

[0085] In the vertical space dimension, the coordinate system uses the positive Z-axis region to represent the ground-level spatial positions such as the vertical height of the vertical cultivation rack and greenhouse. For example, Z=1.8 represents a height of 1.8 meters above ground, corresponding to the third layer of the vertical cultivation rack; Z=3.5 represents a height of 3.5 meters above ground, corresponding to the top area of ​​the greenhouse. Vertical spatial coordinates are of great significance for the management of new planting models such as vertical agriculture and vertical cultivation.

[0086] The time dimension is represented by a T-coordinate, recording the temporal location of the planting unit within its growth cycle. The time coordinate typically uses the ISO 8601 standard time format, such as "2024-03-15T14:30:00" representing 14:30 on March 15, 2024. The introduction of the time coordinate imbues spatial data with temporal attributes, supporting spatiotemporal joint analysis.

[0087] The data structure of dynamic spatial tags includes the following fields: Planting Unit Identifier ID, Multidimensional Spatial Coordinates (X, Y, Z, T), Tag Creation Timestamp, Tag Update Timestamp, Tag Validity Period, and Spatial Status Marker. The Planting Unit Identifier ID is a unique identifier for the tag, used to associate the planting unit's identity information. The Multidimensional Spatial Coordinates record the current spatial location and time of the planting unit. The Tag Creation Timestamp records the time when the tag was first generated. The Tag Update Timestamp records the time of the tag's most recent update. The Tag Validity Period sets the effective time window for the tag; tags exceeding their validity period need to have their data re-collected and updated. The Spatial Status Marker records the current status of the planting unit, such as normal, migrating, abnormal, or harvested.

[0088] The lifecycle management of dynamic spatial tags includes three phases: creation, updating, and expiration. In the creation phase, when a planting unit is first bound to spatial coordinates, an initial dynamic spatial tag is generated, and a tag creation timestamp and validity period are set. In the updating phase, when a change in the spatial location of a planting unit or a change in environmental parameters is detected, the corresponding fields in the dynamic spatial tag are updated, and the tag update timestamp is refreshed. In the expiration phase, when a tag expires without being updated or the planting unit is harvested, the tag is marked as invalid, and data collection and updating cease.

[0089] See Figure 3 The construction process of the spatiotemporal graph is divided into three stages: spatial adjacency graph construction, time series construction, and spatiotemporal fusion.

[0090] 1. Spatial Adjacency Graph Construction Stage A spatial adjacency graph is a graph structure with planting units as nodes and spatial adjacency relationships as edges. The key to constructing a spatial adjacency graph is determining the spatial adjacency relationships between planting units.

[0091] The method for determining spatial adjacency is as follows: First, calculate the spatial distance between two planting units. Spatial distance can be measured using Euclidean distance, Manhattan distance, or other distance metrics. For three-dimensional spatial coordinates, the formula for calculating Euclidean distance is:

[0092] in, The distance between two planting units is represented in meters; X1, Y1, and Z1 represent the X-axis, Y-axis, and Z-axis coordinates of the first planting unit in the multidimensional unified coordinate system, respectively; X2, Y2, and Z2 represent the X-axis, Y-axis, and Z-axis coordinates of the second planting unit in the multidimensional unified coordinate system, respectively.

[0093] Then, the calculated spatial distance is compared with a preset adjacency distance threshold. If the spatial distance is less than the adjacency distance threshold, it is determined that the two planting units have a spatial adjacency relationship, and an edge is added to the graph to connect the two nodes; otherwise, it is determined that the two planting units do not have a spatial adjacency relationship.

[0094] The adjacency distance threshold should be set considering the type of planting unit and planting management needs. For example, for adjacent plants, the adjacency distance threshold can be set to 1.5 times the plant crown width to cover the overlapping area of ​​the plant crowns; for planting containers, the adjacency distance threshold can be set to the container spacing plus a certain margin to identify physically adjacent containers; for vertical cultivation planting troughs with different layers, the adjacency distance threshold can be set to 1.2 times the layer spacing to identify vertically adjacent planting troughs.

[0095] Once the spatial adjacency graph is constructed, it can be used for various spatial analysis applications. For example, it can identify the spatial neighbors of planting units for early warning of pest and disease transmission; calculate the spatial connectivity of planting units for resource scheduling optimization; and discover spatial clusters of planting units for planting layout analysis.

[0096] 2. Time series construction stage A time series is a sequence of spatial state data of planting units at different points in time. Constructing a time series requires organizing the records of spatial state data changes of planting units in chronological order.

[0097] The sources of time-series data include update records of dynamic spatial labels, migration records in spatial trajectory tables, and sampling data from space environment sensors. These data are sorted by timestamps to form time-series data.

[0098] Time series data can be represented as:

[0099] in This represents a time-series data set representing planting units; This indicates the length of the time series, i.e., the number of sampling points; Indicates a point in time The spatial state at that time includes information such as spatial location, environmental parameters, and growth status; , ,..., This represents n sampling time points arranged in chronological order, satisfying t1 <t2<...<t n .

[0100] The sampling frequency of time series data should be determined based on monitoring requirements. For rapidly changing parameters (such as temperature and humidity), the sampling frequency can be higher, such as sampling every 5 minutes; for slowly changing parameters (such as growth stage), the sampling frequency can be lower, such as sampling once a day. Data of different frequencies can be unified to the same time grid through interpolation, aggregation, or other methods.

[0101] 3. Spatiotemporal Fusion Stage Spatiotemporal fusion is the process of fusing spatial adjacency graphs and time-series data to generate spatiotemporal graph node representations. This invention employs a spatiotemporal graph neural network to achieve spatiotemporal fusion.

[0102] The input to the spatiotemporal graph neural network includes a spatial adjacency graph and time series data. The spatial adjacency graph is represented by an adjacency matrix A and a node feature matrix X, where A is an N×N matrix (N is the number of nodes), A[i,j]=1 indicates that nodes i and j are adjacent, and A[i,j]=0 indicates that there is no adjacency relationship; X is an N×D matrix (D is the feature dimension), and X[i,:] represents the feature vector of node i. The time series data is represented in tensor form with dimensions N×T×F, where T is the number of time steps, and F is the feature dimension of each time step.

[0103] The processing flow of the spatiotemporal graph neural network is as follows: First, the spatial convolutional layer performs graph convolution on the spatial adjacency graph. The formula for graph convolution is:

[0104] in, This represents the adjacency matrix with added self-loops. = +I, A is the adjacency matrix of the spatial adjacency graph, I is the identity matrix; D represents the degree matrix of ∠A, which is a diagonal matrix whose diagonal elements D[i,i] are the sum of the elements in the i-th row of the adjacency matrix ∠A; H (l) H represents the feature matrix of the nodes in the l-th layer. (l- ¹ ) H represents the node feature matrix of the (l-1)th layer. (0) Let X be the input node feature matrix; W be the input node feature matrix. (l) Let represent the learnable weight matrix of the l-th layer; σ represents the activation function, such as ReLU, sigmoid, etc.; D - ¹ / ² represents the degree matrix D raised to the power of -1 / 2, i.e., the square root of the inverse of the degree matrix. It is used to normalize the adjacency matrix, keeping the scale of the eigenvalues ​​stable. Graph convolution operations aggregate the feature information of neighboring nodes to generate spatial adjacency features.

[0105] Then, the temporal convolutional layer performs temporal convolution operations on the time series. The temporal convolution operation uses one-dimensional convolution, sliding the convolution kernel along the time dimension to extract local features of the time series. Temporal convolution operations can capture short-term patterns and long-term dependencies in the time series, generating temporally varying features.

[0106] Finally, the spatiotemporal attention layer fuses spatial adjacency features and temporal variation features. The calculation formula for the attention mechanism is:

[0107] Where Q represents the query matrix, obtained through a linear transformation of spatial adjacency features, used to retrieve relevant information; K represents the key matrix, obtained through a linear transformation of time-varying features, used to match the query; and V represents the value matrix, obtained through a linear transformation of time-varying features, used to output the final feature values; QK K represents the transpose of the query matrix Q and the key matrix K. Matrix multiplication is used to calculate the similarity between the query and the key; d k The dimension of the key matrix K is used to scale the dot product result to prevent gradient vanishing or exploding; softmax(·) represents the softmax normalization function, which converts the similarity score into a probability distribution; √d k d k The square root of the product is used to scale the dot product result. The spatiotemporal attention layer can adaptively learn the importance weights of spatial and temporal features, generating spatiotemporal graph node representations that integrate spatial and temporal relationships.

[0108] The spatiotemporal graph node representation is an N×D' matrix, where D' is the dimension of the node representation. The representation vector of each node integrates the spatial adjacency information, temporal evolution information, and feature information of that node, and can be used for downstream spatial analysis tasks.

[0109] The whole-domain spatial planting data management method and system provided in this application can be applied to a variety of agricultural scenarios. The application effect is illustrated below through specific embodiments.

[0110] Application Scenario 1: Spatial Data Management Across Greenhouse and Vertical Agriculture An agricultural technology park operates both a greenhouse cultivation area and a vertical farming area. The greenhouse cultivation area covers 2,000 square meters and grows crops such as tomatoes and cucumbers; the vertical farming area consists of six tiers of vertical cultivation racks, each 30 centimeters high, primarily growing leafy vegetables. The park needs to achieve unified management of spatial planting data for both areas and support cross-regional spatial queries and analysis.

[0111] Using the method and system provided by this invention, the park has established a unified multi-dimensional spatial coordinate system, mapping the spatial data of the greenhouse planting area and the vertical agricultural planting area to the same coordinate system. The planting units in the greenhouse planting area are divided by ground grids, with each grid measuring 5 meters × 5 meters, for a total of 100 planting units; the planting units in the vertical agricultural planting area are divided by planting troughs, for a total of 288 planting units (6 layers × 4 columns × 12 troughs).

[0112] The system deploys distributed spatial sensing devices, including environmental sensors and positioning equipment, to collect real-time spatial status data of each planting unit. Based on the collected data, the system dynamically updates the dynamic spatial tags of the planting units and records the trajectory of spatial location changes.

[0113] For cross-scenario data fusion, the system standardizes the data from greenhouse and vertical agriculture scenarios and stores them in a unified database. The system constructs a spatiotemporal map to link planting units in the two regions, supporting cross-regional spatial analysis.

[0114] Park managers can use the system to query planting units and their status within any spatial range, trace the spatial trajectory of any planting unit, analyze pest and disease transmission paths, and predict growth trajectories. Cross-scenario data fusion enables managers to uniformly plan planting resources in two areas, improving management efficiency.

[0115] Application Scenario 2: Spatial Tracking and Decision Support for Field Planting A large farm covers an area of ​​5,000 mu (approximately 333 hectares) and grows crops such as wheat and corn. The farm employs mechanized operations, and crops may require regional adjustments at different growth stages (such as replanting or transplanting). The farm needs to track the crop growth status in each area and support precision agriculture decision-making.

[0116] Using the method and system provided by this invention, the farm has established a multi-dimensional spatial coordinate system covering the entire farm. Due to the large planting area, the coordinate system adopts a combination of geographic coordinates (latitude and longitude) and local coordinates to support large-scale spatial positioning.

[0117] The planting units are divided into management units, each approximately 10 acres, totaling 500 planting units. The system generates a dynamic spatial tag for each planting unit, recording its spatial location and boundaries.

[0118] The system deploys IoT sensors and remote sensing equipment to collect spatial environmental data and crop growth data from each management unit. The system regularly updates dynamic spatial tags to record changes in the status of the planting units.

[0119] The system constructs a spatiotemporal map, linking the spatial relationships and temporal evolution of each management unit. Based on the spatiotemporal map, the system can perform spatial clustering analysis to identify regions with similar growth states; perform spatial hotspot analysis to discover abnormal areas; and predict growth trajectories to estimate harvest time and yield.

[0120] Using the method and system provided by this invention, a multi-dimensional spatial coordinate system was established at the base, dividing the greenhouse space into multiple monitoring units. The system collects spatial environmental data and pest and disease monitoring data from each monitoring unit in real time.

[0121] When pests and diseases are detected, the system analyzes their spread based on a spatiotemporal map. The spatiotemporal map records the spatial adjacency relationships and temporal evolution of each monitoring unit, allowing the system to trace the spread of pests and diseases from their initial location to their current location.

[0122] The system achieves pest and disease transmission tracing through the following steps: First, it identifies the monitoring unit where the pest or disease first appeared and determines the initial time point based on the time sequence of pest and disease monitoring data; then, it traces the path of the pest or disease from the initial location to adjacent units based on spatial adjacency; finally, it outputs a visualization of the transmission path, indicating the transmission direction and time.

[0123] Base managers can quickly locate the source of pests and diseases and take targeted prevention and control measures; they can predict potentially affected areas based on the transmission path and take preventive measures in advance; and they can evaluate the effectiveness of prevention and control measures and adjust prevention and control strategies accordingly.

[0124] The whole-domain spatial planting data management method and system provided by this invention achieves the following technical effects through technical means such as establishing a multi-dimensional unified spatial coordinate system, designing a dynamic spatial labeling mechanism, realizing cross-scene data fusion, and constructing a spatiotemporal map: (1) Unified management of multi-dimensional space. By using a quadruple coordinate system, the horizontal space, underground space, vertical space and time dimension of the ground are managed in a unified manner, eliminating spatial data silos and realizing integrated query and analysis of multi-dimensional spatial data.

[0125] (2) Dynamic spatial tracking. The dynamic spatial tagging mechanism enables dynamic tracking of the spatial location of planting units, records the spatial trajectory of planting units throughout the entire growth cycle, and supports spatial tracing and process analysis.

[0126] (3) Cross-scene data fusion. Through scene recognition, coordinate system transformation and data format standardization, spatial data fusion of different agricultural scenes such as greenhouses, open fields, vertical agriculture and rooftop planting has been achieved, breaking down data barriers between scenes.

[0127] (4) Spatiotemporal joint analysis. The spatiotemporal graph and spatiotemporal graph neural network realize the joint modeling of time and space dimensions, and support spatiotemporal joint analysis applications such as tracing the spread of pests and diseases and predicting growth trajectory.

[0128] (5) Intelligent decision support. Based on the spatiotemporal map analysis results, suggestions for optimizing planting layout, prediction results of growth trajectory and resource scheduling scheme can be generated, providing intelligent support for planting management decisions.

[0129] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application. These modifications or substitutions should all be considered to be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for managing planting data in a global space, characterized by, Includes the following steps: Establish a multi-dimensional unified spatial coordinate system for the entire planting area. The multi-dimensional unified spatial coordinate system covers the horizontal space, underground space and vertical space of the entire planting area, and assigns a unique multi-dimensional spatial coordinate identifier to each spatial location within the entire planting area. The planting unit is bound to the multidimensional spatial coordinate identifier corresponding to the planting unit in the multidimensional spatial unified coordinate system, and a dynamic spatial label including the binding relationship between the planting unit and its corresponding multidimensional spatial coordinate identifier is generated. Spatial status data of each planting unit is collected by distributed spatial sensing devices. The multi-dimensional spatial coordinate identifier in the dynamic spatial label of the planting unit is updated according to the spatial status data, and the spatial position change trajectory of the planting unit is recorded. Acquire spatial planting data under different agricultural scenarios, and standardize the spatial planting data based on the multi-dimensional unified spatial coordinate system to obtain standardized spatial planting data. A global spatial planting database is constructed based on the dynamic spatial labels, the spatial state data, the spatial location change trajectory, and the standardized spatial planting data. A spatiotemporal map is constructed based on the global spatial planting database, and spatial analysis results of planting units are output based on the spatiotemporal map.

2. The method of claim 1, wherein, The establishment of a unified coordinate system in multidimensional space includes: Multidimensional spatial coordinates are represented by a quadruple: (X, Y, Z, T), where X and Y represent the horizontal coordinates on the ground, Z represents the vertical coordinates, and the positive and negative values ​​of Z represent the vertical height and underground depth, respectively, and T represents the time coordinate. The entire planting area is divided into multiple spatial grid units according to a preset spatial resolution, and a unique spatial grid code is assigned to each spatial grid unit.

3. The method of claim 1, wherein, The step of binding the planting unit with its corresponding multidimensional spatial coordinate identifier in a unified multidimensional spatial coordinate system, and generating a dynamic spatial label that includes the binding relationship between the planting unit and its corresponding multidimensional spatial coordinate identifier, includes: Assign a unique identifier to the planting unit and obtain the initial multidimensional spatial coordinate identifier of the planting unit; The unique identifier of the planting unit is bound to the initial multidimensional spatial coordinate identifier to generate an initial dynamic spatial label and store it.

4. The method according to claim 1, characterized in that, The process of collecting spatial status data of each planting unit through distributed spatial sensing devices includes: The horizontal coordinates of the planting unit are collected using a ground position sensor; The vertical height or underground depth coordinates of the planting unit are collected by a vertical position sensor; Temperature, humidity, light, or nutrient concentration parameters of the space where the planting unit is located are collected by spatial environment sensors; The spatial adjacency relationship of the planting units is determined based on the spatial distance between each planting unit.

5. The method according to claim 3, characterized in that, Update the multidimensional spatial coordinate identifier in the dynamic spatial label of the planting unit according to the spatial state data, and record the spatial position change trajectory of the planting unit, including: Compare the current multidimensional spatial coordinate identifier of the planting unit with the historical coordinate identifier, and calculate the spatial position change. The historical coordinate identifier includes the initial multidimensional spatial coordinate identifier of the planting unit. When the change in spatial location exceeds a preset threshold, the multidimensional spatial coordinate identifier of the planting unit is updated and a migration record is added to the spatial trajectory table. When the change does not exceed the threshold, the environmental parameters in the dynamic spatial label are updated according to the changes in environmental parameters.

6. Obtain the current multidimensional spatial coordinate identifier of the planting unit, compare the current multidimensional spatial coordinate identifier with the historical multidimensional spatial coordinate identifier recorded in the dynamic spatial label, and calculate the spatial position change; The location change threshold is determined based on the type of planting unit and spatial migration characteristics. When the spatial location change exceeds the location change threshold, it is determined that a spatial migration event has occurred. A location record is added to the spatial trajectory table, and the multi-dimensional spatial coordinate identifier in the dynamic spatial label is updated. When the change in spatial location does not exceed the location change threshold, the change in spatial environment data is acquired, and the environmental parameters in the dynamic spatial tag are updated based on the change in spatial environment data.

7. The method according to claim 1, characterized in that, The process of acquiring spatial planting data under different agricultural scenarios, and standardizing the spatial planting data based on the multi-dimensional unified spatial coordinate system to obtain standardized spatial planting data, includes: Acquire spatial planting data, and identify agricultural scene types based on the spatial planting data. The agricultural scene types include greenhouse scenes, field scenes, vertical agriculture scenes, or rooftop planting scenes. Based on the agricultural scenario type, obtain the corresponding coordinate system transformation rules and convert the local coordinate system of the spatial planting data into the multidimensional spatial coordinate values ​​of the multidimensional unified spatial coordinate system. The converted spatial planting data is subjected to spatial consistency verification, and the spatial planting data that passes the verification is stored in the global spatial planting database.

8. The method according to claim 1, characterized in that, The construction of the spatiotemporal map based on the global spatial planting database includes: Calculate the spatial distance between each planting unit. When the spatial distance is less than a preset adjacency distance threshold, it is determined that two planting units have a spatial adjacency relationship. A spatial adjacency graph structure is constructed using the planting units as nodes and the spatial adjacency relationships between the planting units as edges; Organize the spatial status data change sequence of each planting unit according to time sequence to construct time series data; The spatial adjacency graph structure and the time series data are input into the spatiotemporal graph neural network, and the spatiotemporal graph node representation is output.

9. The method according to claim 7, characterized in that, The spatiotemporal graph neural network includes spatial convolutional layers, temporal convolutional layers, and spatiotemporal attention layers: The spatial convolutional layer is used to perform graph convolution operations on the spatial adjacency graph structure to extract the spatial adjacency features of the planting units. The temporal convolutional layer is used to perform temporal convolution operations on the time series data to extract the temporal variation features of the planting units. The spatiotemporal attention layer is used to fuse the spatial adjacency features and the temporal variation features, and output a spatiotemporal graph node representation that fuses spatial and temporal relationships.

10. The method according to claim 1, characterized in that, The spatial analysis results of the planting units output based on the spatiotemporal map include: Based on the aforementioned spatiotemporal map, the transmission path of pests and diseases can be traced according to the temporal sequence and spatial location of their occurrence. Based on the spatiotemporal map, the spatial location of the planting unit at future time points is predicted according to the historical spatial trajectory sequence and current spatial state data of the planting unit, and the growth trajectory prediction result is generated.

11. A comprehensive spatial planting data management system, characterized in that, include: The multi-dimensional space coordinate management module is used to establish and manage the aforementioned unified multi-dimensional space coordinate system. The planting unit spatial binding module is used to bind planting units to multi-dimensional spatial coordinate identifiers and generate the dynamic spatial labels; A distributed spatial sensing layer is used to collect spatial status data of each planting unit; The dynamic spatial label management module is used to update the multi-dimensional spatial coordinate identifiers in the dynamic spatial labels and record the trajectory of spatial location changes. The cross-scenario data standardization module is used to standardize spatial planting data in different agricultural scenarios; The global spatial planting database construction module is used to construct the global spatial planting database; The spatiotemporal map construction module is used to construct a spatiotemporal map based on the global spatial planting database; The spatial analysis module is used to output the spatial analysis results of the planting units based on the spatiotemporal map; The system is configured to perform the method according to any one of claims 1 to 9.

Images