CN117907242A - Homeland mapping method, system and storage medium based on dynamic remote sensing technology - Google Patents
Homeland mapping method, system and storage medium based on dynamic remote sensing technology Download PDFInfo
- Publication number
- CN117907242A CN117907242A CN202410294927.4A CN202410294927A CN117907242A CN 117907242 A CN117907242 A CN 117907242A CN 202410294927 A CN202410294927 A CN 202410294927A CN 117907242 A CN117907242 A CN 117907242A
- Authority
- CN
- China
- Prior art keywords
- data
- flight
- ground control
- mapping
- control point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013507 mapping Methods 0.000 title claims abstract description 86
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000005516 engineering process Methods 0.000 title claims abstract description 33
- 230000008859 change Effects 0.000 claims abstract description 32
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 19
- 230000002787 reinforcement Effects 0.000 claims abstract description 12
- 238000005457 optimization Methods 0.000 claims abstract description 11
- 239000003795 chemical substances by application Substances 0.000 claims description 23
- 230000011218 segmentation Effects 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 238000012544 monitoring process Methods 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 10
- 238000003709 image segmentation Methods 0.000 claims description 8
- 230000007613 environmental effect Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 239000003344 environmental pollutant Substances 0.000 claims description 3
- 230000007774 longterm Effects 0.000 claims description 3
- 231100000719 pollutant Toxicity 0.000 claims description 3
- 230000001502 supplementing effect Effects 0.000 claims description 3
- 238000005260 corrosion Methods 0.000 claims description 2
- 230000007797 corrosion Effects 0.000 claims description 2
- 238000005286 illumination Methods 0.000 claims description 2
- 239000000463 material Substances 0.000 claims description 2
- 238000001556 precipitation Methods 0.000 claims description 2
- 239000002243 precursor Substances 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 239000013618 particulate matter Substances 0.000 claims 1
- 230000008569 process Effects 0.000 description 8
- 238000010422 painting Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 230000000875 corresponding effect Effects 0.000 description 4
- 238000007726 management method Methods 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013209 evaluation strategy Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 238000002834 transmittance Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to the technical field of dynamic remote sensing mapping, and particularly discloses a homeland mapping method, a homeland mapping system and a storage medium based on a dynamic remote sensing technology, wherein the method comprises the following steps: dividing a region to be painted into different subregions, determining the positions of ground control points by using a convolutional neural network and reinforcement learning, generating an optimal path laid by the ground control points of each subregion by using an optimization algorithm, detecting real-time environment data of each subregion, constructing a flight mapping knowledge graph, deducing whether a flight environment can execute a flight plan by using the knowledge graph, determining unmanned aerial vehicle and camera parameters which can be used by the flight environment, carrying out regular flight mapping on the region to be mapped, and determining the change condition by comparing flight mapping results in different periods. The application combines dynamic remote sensing technology, convolutional neural network, reinforcement learning and optimization algorithm, can efficiently and accurately survey and draw the territory, not only improves the automation degree of surveying and drawing, but also remarkably improves the accuracy and efficiency of surveying and drawing.
Description
Technical Field
The invention relates to the technical field of dynamic remote sensing mapping, in particular to a homeland mapping method, a homeland mapping system and a storage medium based on a dynamic remote sensing technology.
Background
The homeland mapping system based on the dynamic remote sensing technology has wide application in the remote sensing field, however, for real-time monitoring and judging of environmental factors, especially the integration of air environment information, which is not fully paid attention to in the field so far, the current homeland mapping system has a problem that the change of the environmental factors cannot be timely focused and evaluated in the process of mapping, the change comprises the fluctuation of atmospheric pollution and meteorological conditions, the factors possibly directly influence the acquisition and processing of remote sensing data of flying equipment, thereby reducing the accuracy and quality of mapping results, and the lack of real-time monitoring of the air environment means that people cannot adjust at key moments so as to ensure that the acquired data is reliable and accurate.
The prior art publication No. CN117493818A provides a homeland mapping method, a system and a storage medium based on a dynamic remote sensing technology, wherein the system acquires and processes air environment information in a flight area in real time through an environment acquisition module, a weather acquisition module, a preprocessing module, an analysis module, an evaluation module and a decision module to acquire a flight painting monitoring index Jczs, matches a flight painting monitoring threshold F preset by the evaluation module with the flight painting monitoring index Jczs to acquire a flight painting evaluation strategy scheme, and finally specifically executes the content of the flight painting evaluation strategy scheme through the decision module. The air environment information can be monitored in real time, so that the system can timely sense the changes of atmospheric pollution and meteorological conditions in the flight drawing and measuring process, data acquisition is facilitated under favorable conditions, the influence of adverse environmental factors in the drawing and measuring process is reduced, and the accuracy and quality of remote sensing data are improved.
The above prior art solutions, although realizing the relevant beneficial effects by the prior art structure, still have the following drawbacks: mapping is carried out only through aerial images, so that the problems of low mapping precision, information loss and the like are caused, whether a flight plan can be executed is judged through a single threshold value, misjudgment is easily caused, the flight environment cannot be accurately estimated, the position of a ground control point in the prior art is determined mainly by an expert, and higher professional literacy is required and the efficiency is lower.
Therefore, the invention provides a homeland mapping method, a homeland mapping system and a storage medium based on a dynamic remote sensing technology, and aims to solve the problems.
Disclosure of Invention
1. The technical problem to be solved.
The application aims to provide a territorial mapping method, a territorial mapping system and a storage medium based on a dynamic remote sensing technology, which solve the technical problems in the background technology and realize the determination of the position of a ground control point through a convolutional neural network and reinforcement learning; the optimal path for ground control point layout is generated through an optimization algorithm, mapping efficiency is improved, and air environment data of each sub-area are monitored in real time through the arrangement of the air environment data sensor, so that an important basis is provided for adjustment of a flight plan.
2. The technical proposal is that.
The invention provides a homeland mapping method, a homeland mapping system and a storage medium based on a dynamic remote sensing technology, which are used for improving the efficiency and the accuracy of homeland mapping.
The application provides a homeland mapping method based on a dynamic remote sensing technology, which comprises the following steps.
S1, dividing the region to be painted into different subareas according to the route, the flight altitude, the flight speed and the flight plan content which are set by software assistance.
S2, determining the position of the ground control point by using a convolutional neural network and reinforcement learning, and generating an optimal path laid by the ground control point of each sub-area by using an optimization algorithm.
S3, setting an air environment data sensor in each sub-area for detecting real-time environment data of each sub-area.
S4, constructing a knowledge graph by combining air environment data and weather report data according to the model and camera parameters of the unmanned aerial vehicle.
S5, using the knowledge graph to infer whether the flight environment can execute the flight plan, determining unmanned aerial vehicle models and camera parameters which can be used by the flight environment, and the like.
S6, carrying out regular flight mapping on the area to be mapped, and comparing flight mapping results in different periods to determine the change condition.
As an alternative of the present invention, in step S2, the position determining method of the ground control point includes the following steps.
S201, preparing a large number of marked aerial image data sets, wherein each pixel is marked with a corresponding ground object type label, and dividing the aerial image data into a training set and a testing set. Satellite images of areas requiring homeland mapping are collected.
S202, inputting the aerial images of the training set into a DeepLabv & lt3+ & gt model constructed by using Pytorch frames for training, and updating model weights through a back propagation algorithm, so that the model can accurately predict the ground object category to which each pixel belongs.
And S203, after training convergence, evaluating the performance of the model by using a test set, and carrying out pixel-level semantic segmentation prediction on the satellite image of the region needing to be subjected to homeland mapping to obtain a DeepLabv3+ segmentation map.
S204, flattening the DeepLabv3+ split map, namely converting the three-dimensional tensor of H multiplied by W multiplied by C (height, width and channel number) into a one-dimensional vector form with the length of HWC, and taking the one-dimensional vector form as the state input of the intelligent agent.
S205, defining an agent, wherein the output format of the agent is the actual coordinate (x, y) set of the possible control points, and the two coordinate values of the possible control points can be continuously changed within the allowable range.
S206, designing a reward function: after each step of selection, calculating scores of possible control points selected at the time in the aspects of geometric intensity, definition, elevation and the like, and giving a reward value.
S207, selecting an optimal control point by using the deep Q learning DQN through continuous trial and error learning as a ground control point selection strategy.
As an alternative of the present invention, in step S2, the convolutional neural network CNN is used to perform semantic segmentation on the aerial image, divide the image into different categories, extract the ground feature information of the image, provide a reference for the subsequent ground control point selection, and train an intelligent body using RL, where the objective of the intelligent body is to select the ground control points on the image, so that the number, distribution, quality, etc. of the ground control points meet a certain standard. The input of the intelligent agent is the semantic segmentation result of the image, and the output is the position of the ground control point. The rewarding function of the intelligent agent is designed according to the geometric intensity, the target definition, the elevation fluctuation and other factors of the ground control points so as to encourage the intelligent agent to select proper ground control points, the intelligent agent is used for selecting the ground control points on a new aerial image, the positions and the numbers of the ground control points are output, and an automatic ground control point selecting task is completed.
As an alternative scheme of the invention, the optimal path determination method for the ground control points of each sub-area is that the optimal path of each sub-area is calculated by using Dijkstra algorithm according to the design position abstraction of the ground control points as a directed graph and the ground control points as nodes and the reachability between the ground control points as edges.
As an alternative of the present invention, in step S4, the knowledge graph construction method includes the following steps.
S41, acquiring real-time environment data, weather report data of each sub-area, callable unmanned aerial vehicle model and camera parameter information, storing the information as three data tables, preprocessing the data, and adding a main key for each data table.
S42, identifying the entities in the knowledge graph from the three data tables, adding a label for each entity to represent the entity type, and adding a name for each entity to represent the specific name of the entity.
S43, extracting the relation among the entities from the three data tables, assigning a type to each relation, representing the nature of the relation, assigning a direction to each relation, and representing the direction of the relation.
S44, constructing an ontology according to the types of the entities and the relations, defining the attribute of each entity class, defining the attribute of each relation type, and defining the constraint condition of each entity class and each relation type.
S45, storing the entity, the relation and the attribute in a graph database to form a knowledge base of a graph structure, adding information such as identifiers, labels, names, attributes and constraint conditions for each node (entity) and each side (relation), and adding an index for the knowledge base, so that data can be conveniently and rapidly queried and retrieved.
S46, reasoning the data in the knowledge base by utilizing the ontology and the logic rules, supplementing missing information, finding new knowledge, improving the quality and the integrity of the knowledge, reasoning the evaluation result of the flight environment of each sub-area according to the influence relationship of air environment data and weather report data on the flight environment, reasoning the camera parameters which can be used by each unmanned plane model according to the matching relationship of the unmanned plane model and the camera parameters, and storing the information in the knowledge base as the relationship and the attribute.
S47, inquiring an evaluation result of the flight environment in the knowledge graph according to the flight plan input by the user, judging whether the flight plan can be executed, if so, inquiring a matching relation between the unmanned aerial vehicle model and the camera parameters in the knowledge graph, selecting proper unmanned aerial vehicle model and camera parameters, making the flight plan, and feeding back the information to the user.
As an alternative of the present application, in step S6, the types of the pixels of the flight mapping image at different periods are compared by image segmentation, the change of the surface type, the spatial distribution condition and the amount of change are determined, and the change of the terrain is determined by periodically measuring the position information of the ground control point.
As an alternative scheme of the application, the application provides a homeland mapping system based on a dynamic remote sensing technology, which comprises the following components.
Control point layout module: the module uses convolutional neural network and reinforcement learning to determine the position of ground control points, and uses an optimization algorithm to generate an optimal path for the ground control points of each sub-area. Therefore, the control point layout speed can be increased and the mapping efficiency can be improved in the mapping process.
The subarea dividing module: and dividing the region to be painted into different subareas according to the set flight altitude, flight speed and flight plan content of the route. Thus, the whole mapping area can be decomposed into smaller and easier to manage parts, and subsequent data processing and analysis are convenient.
The environmental data monitoring module: an air environment data sensor is arranged in each sub-area and is used for detecting air environment data of each sub-area, such as temperature, humidity, wind speed, wind direction, atmospheric pollutant concentration and the like in real time. These data may provide an important basis for subsequent flight plan adjustments.
Knowledge graph construction module: and combining the called unmanned aerial vehicle model, camera parameters, and air environment data and weather report data acquired in real time to construct a knowledge graph. Knowledge maps can help the system to better understand and utilize the data, and improve the accuracy of decision making.
Flight environment assessment module: and utilizing the knowledge graph to infer whether the flight environment can execute the flight plan, and determining the usable unmanned plane model, camera parameters and the like. Therefore, flight mapping can be ensured under proper conditions, and the accuracy of mapping results is improved.
The aerial survey execution module: and carrying out periodical aerial survey on the region to be mapped, and judging the change of the earth surface type, the spatial distribution condition and the change quantity by using an image segmentation technology. Thus, the latest surface information can be obtained, and accurate data support is provided for homeland mapping.
And the data processing and analyzing module is used for: and the system is responsible for collecting and processing data generated by each module, and the change trend and the space distribution condition of the earth surface type are revealed by analyzing and comparing mapping results of different time points, so that decision basis is provided for homeland planning and resource management.
And a storage and output module: the processed data is stored in a proper storage medium, such as a hard disk, cloud storage and the like, and an interface for outputting the data is provided for other systems or personnel to review and use. Meanwhile, visual products such as reports or charts can be generated, and the visual products are convenient for non-professional staff to understand and use.
As an alternative of the present application, the present application provides a computer-readable storage medium having stored thereon a computer program that, when run, executes a control point layout module, a sub-region division module, a data processing and analysis module, a storage and output module, a flight environment assessment module, a navigation survey execution module, an environment data monitoring module, and a knowledge graph construction module.
3. Has the beneficial effects of.
One or more of the technical solutions provided in the technical solution of the present invention have at least the following technical effects or advantages.
1. Determining the position of a ground control point through a convolutional neural network and reinforcement learning; and an optimal path for ground control point layout is generated through an optimization algorithm, so that mapping efficiency is improved.
2. The change of the earth surface type, the space distribution condition and the change quantity can be rapidly obtained through the periodical aerial survey and image segmentation technology, and the mapping efficiency is further improved.
3. By arranging the air environment data sensor, the air environment data of each sub-area is monitored in real time, and an important basis is provided for adjustment of a flight plan.
4. The unmanned aerial vehicle model, the camera parameters, the air environment data and the weather report data are combined to construct a knowledge graph, and whether the flight environment can execute a flight plan is inferred through the knowledge graph, so that flight mapping is ensured under proper conditions, and the accuracy of mapping results is improved.
5. Through the data processing and analyzing module, the change trend and the space distribution condition of the earth surface type are revealed, decision basis is provided for homeland planning and resource management, understanding and use by non-professional staff are facilitated, and decision support capability is improved.
Drawings
Fig. 1 is a flow chart of a homeland mapping method based on a dynamic remote sensing technology.
Fig. 2 is a flow chart of a method of determining the position of a ground control point.
Fig. 3 is a flowchart of an optimal path determination method for the layout of sub-area ground control points.
Fig. 4 is a flowchart of a knowledge graph construction method.
Fig. 5 is a diagram of a homeland mapping system based on a dynamic remote sensing technology according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the following detailed description of the embodiments of the present application will be given with reference to the accompanying drawings.
Referring to fig. 1, the present application provides a homeland mapping method based on dynamic remote sensing technology, which comprises the following steps.
S1, dividing the region to be painted into different subareas according to the set flight altitude, flight speed and flight plan content of the route.
S2, determining the position of the ground control point by using a convolutional neural network and reinforcement learning, and generating an optimal path laid by the ground control point of each sub-area by using an optimization algorithm.
S3, setting an air environment data sensor in each sub-area for detecting real-time environment data of each sub-area.
S4, constructing a knowledge graph by combining air environment data and weather report data according to the model and camera parameters of the unmanned aerial vehicle.
S5, using the knowledge graph to infer whether the flight environment can execute the flight plan, determining unmanned aerial vehicle models and camera parameters which can be used by the flight environment, and the like.
S6, carrying out regular flight mapping on the area to be mapped, and comparing flight mapping results in different periods to determine the change condition.
In step S6, the image segmentation model DeepLabv3+ is used to compare the types of the pixels of the flight mapping images in different periods, determine the change, the spatial distribution condition and the change amount of the earth surface type (including mountain, river, grassland, vegetation, etc.), deepLabv3+ is a semantic segmentation model based on deep learning, and can classify each pixel of the image, determine the change, the spatial distribution condition and the change amount of the earth surface type by the change of the types of the pixels of the flight mapping images in different periods, and determine the change of the topography by periodically measuring the position information of the ground control points.
Referring to fig. 2, in step S2, the method for determining the position of the ground control point includes the following steps.
S201, preparing a large number of marked aerial image data sets, wherein each pixel is marked with a corresponding ground object type label, and dividing the aerial image data into a training set and a testing set. Satellite images of areas requiring homeland mapping are collected.
S202, inputting the aerial images of the training set into a DeepLabv & lt3+ & gt model constructed by using Pytorch frames for training, and updating model weights through a back propagation algorithm, so that the model can accurately predict the ground object category to which each pixel belongs.
And S203, after training convergence, evaluating the performance of the model by using a test set, and carrying out pixel-level semantic segmentation prediction on the satellite image of the region needing to be subjected to homeland mapping to obtain a DeepLabv3+ segmentation map.
S204, flattening the DeepLabv3+ split map, namely converting the three-dimensional tensor of H multiplied by W multiplied by C (height, width and channel number) into a one-dimensional vector form with the length of HWC, and taking the one-dimensional vector form as the state input of the intelligent agent.
S205, defining an agent, wherein the output format of the agent is the actual coordinate (x, y) set of the possible control points, and the two coordinate values of the possible control points can be continuously changed within the allowable range.
S206, designing a reward function: after each step of selection, calculating scores of possible control points selected at the time in the aspects of geometric intensity, definition, elevation and the like, and giving a reward value.
S207, selecting an optimal control point by using the deep Q learning DQN through continuous trial and error learning as a ground control point selection strategy.
In this embodiment, the environment is an aerial image and a corresponding semantic segmentation result, the agent interacts with the environment, a control point is selected on the image, then the environment returns rewards and new state information, the main network is a deep neural network, a Convolutional Neural Network (CNN) structure is adopted to process image input, the input is a feature vector subjected to CNN semantic segmentation or a vector representation after some preprocessing of an original image, the output is a Q value estimation for each possible ground control point candidate position, namely, an expected value of a long-term accumulated rewards obtained after corresponding actions are performed, the expected value is used for storing experience information (state, action, rewards and next state) generated in the interaction process of the agent and the environment, the agent randomly extracts batches therefrom for learning, the sequence correlation among data is broken, the sample utilization rate is improved, the target network is identical to the main network structure, the target network is a delay version of main network parameters, the target network is used for estimating future discount rewards when calculating Q learning targets, a relatively stable Q value target is provided, the agent is used for avoiding excessive fluctuation, the agent is used for calculating the experience target value of the Q target network from each main body, experience value is taken out from a buffer network, and experience target value is calculated at the same time. The parameters of the subject network are updated by minimizing the gap (mean square error) between the two.
In this embodiment, the reward function determines the selection of the ground control point through three factors of geometric intensity (G), target definition (C) and elevation fluctuation (E), and introduces a weight vector w= [ W G,wC,wE ], which respectively corresponds to the importance of the three factors, where the reward function may be expressed as: r (x, y) =w GG(x,y)+wCC(x,y)+wE E (x, y), where G (x, y) represents the geometric intensity of the ground control point selected at the location (x, y), measures the uniformity of distribution of the newly selected point relative to the selected ground control point, G (x, y) =1/d avg(x,y),davg (x, y) is the average distance from the location (x, y) to the selected ground control point, C (x, y) represents the target sharpness at the location (x, y), C (x, y) =f clarity(segmentation_map[y,x]),fclarity is a mapping function to convert the semantic segmentation label into a sharpness score, E (x, y) represents the elevation relief at the location (x, y), E (x, y) = -elementjdata [ x, y ] - μ, if elementjdata exists, μ is the average elevation of the selected ground control point, and the absolute difference may reflect the elevation relief degree.
Referring to fig. 3, the present application provides a method for determining an optimal path for each sub-area ground control point layout, which includes the following steps.
S208, regarding each ground control point as one node in the graph, and adding a directed edge between two ground control points if reachability exists between the two ground control points, and giving the edge a weight which represents the distance or other cost of moving from one point to the other.
S209, the distance of each node is set to infinity (indicating that it has not been accessed), and the distance of the start point is set to 0.
S210, using a minimum heap method, taking node distance as a sequencing basis, and initially only comprising a starting point.
S211, taking out the node with the smallest currently known distance from the priority queue in each round, checking all adjacent nodes of the node, and updating the distance value. If the distance value is updated, the node is marked as accessed and added to the priority queue.
S212, stopping the algorithm when the priority queue is empty or the currently fetched node is the target node.
S213, by backtracking the 'precursor node' of each node, the shortest path from the starting point to the end point can be obtained.
When the ground control points are distributed, the corrosion-resistant materials are used for distribution, so that the regular mapping and long-term use are facilitated.
Referring to fig. 4, in step S4, the knowledge graph construction method includes the following steps.
S41, acquiring real-time environment data, weather report data of each sub-area, callable unmanned aerial vehicle model and camera parameter information, storing the information as three data tables, preprocessing the data, and adding a main key for each data table.
S42, identifying entities in the knowledge graph from the three data tables, such as subareas, unmanned aerial vehicle models, camera parameters, air environment data, weather report data and the like, endowing each entity with a unique identifier, such as subarea 1, unmanned aerial vehicle model A, camera parameters B and the like, adding a label to each entity, such as subarea, unmanned aerial vehicle model, camera parameters, data and the like, representing the entity type, adding a name to each entity, and representing the specific name of the entity.
S43, extracting relations among entities from three data tables, such as a containing relation of subareas and air environment data, an influence relation of the air environment data and a flight environment, matching relation of unmanned aerial vehicle models and camera parameters, assigning a type to each relation, representing the properties of the relation, such as containing, matching, influence and the like, assigning a direction to each relation, representing the direction of the relation, such as from the subareas to the air environment data and the like.
S44, constructing an ontology according to the types of the entities and the relations, namely, a framework describing the knowledge and the concepts of the specific field, defining the types of the entities, the attributes and the relations, defining the attributes of the entity, such as the number, the name, the area and the like of the subareas, defining the attributes of the entity, such as the positive and negative of the influence relation and the strength and the like, for each relation type, and defining the constraint conditions of the entity, such as that the subareas only comprise one air environment data, one weather report data and the like.
S45, storing the entity, the relation and the attribute in a graph database to form a knowledge base of a graph structure, adding information such as identifiers, labels, names, attributes and constraint conditions for each node (entity) and each side (relation), and adding an index for the knowledge base, so that data can be conveniently and rapidly queried and retrieved.
S46, reasoning the data in the knowledge base by utilizing the ontology and the logic rules, supplementing missing information, finding new knowledge, improving the quality and the integrity of the knowledge, reasoning the evaluation result of the flight environment of each sub-area according to the influence relationship of air environment data and weather report data on the flight environment, reasoning the camera parameters which can be used by each unmanned plane model according to the matching relationship of the unmanned plane model and the camera parameters, and storing the information in the knowledge base as the relationship and the attribute.
S47, inquiring an evaluation result of a flight environment in the knowledge graph according to the flight plan input by the user, judging whether each sub-area covered by the route can execute the flight plan, if so, inquiring a matching relation between the unmanned aerial vehicle model and the camera parameters in the knowledge graph, selecting the proper unmanned aerial vehicle model and the camera parameters, making the flight plan, and feeding back the information to the user.
The real-time environment data comprise wind speed, wind direction, humidity, particle concentration, illumination intensity and light transmittance, the unmanned plane model information comprises model, weight, endurance time, load and flying speed, the working environment range, the camera parameter information comprises model, resolution, focal length and aperture of a camera, and the weather report data comprise weather, air pressure, visibility and precipitation.
Referring to fig. 5, the present application provides a homeland mapping system based on dynamic remote sensing technology, which comprises.
Control point layout module: the module uses convolutional neural network and reinforcement learning to determine the position of ground control points, and uses an optimization algorithm to generate an optimal path for the ground control points of each sub-area. Therefore, the control point layout speed can be increased and the mapping efficiency can be improved in the mapping process.
The subarea dividing module: and dividing the region to be painted into different subareas according to the set flight altitude, flight speed and flight plan content of the route. Thus, the whole mapping area can be decomposed into smaller and easier to manage parts, and subsequent data processing and analysis are convenient.
The environmental data monitoring module: an air environment data sensor is arranged in each sub-area and is used for detecting air environment data of each sub-area, such as temperature, humidity, wind speed, wind direction, atmospheric pollutant concentration and the like in real time. These data may provide an important basis for subsequent flight plan adjustments.
Knowledge graph construction module: and combining the called unmanned aerial vehicle model, camera parameters, and air environment data and weather report data acquired in real time to construct a knowledge graph. Knowledge maps can help the system to better understand and utilize the data, and improve the accuracy of decision making.
Flight environment assessment module: and utilizing the knowledge graph to infer whether the flight environment can execute the flight plan, and determining the usable unmanned plane model, camera parameters and the like. Therefore, flight mapping can be ensured under proper conditions, and the accuracy of mapping results is improved.
The aerial survey execution module: and carrying out periodical aerial survey on the region to be mapped, and judging the change of the earth surface type, the spatial distribution condition and the change quantity by using an image segmentation technology. Thus, the latest surface information can be obtained, and accurate data support is provided for homeland mapping.
And the data processing and analyzing module is used for: and the system is responsible for collecting and processing data generated by each module, and the change trend and the space distribution condition of the earth surface type are revealed by analyzing and comparing mapping results of different time points, so that decision basis is provided for homeland planning and resource management.
And a storage and output module: the processed data is stored in a proper storage medium, such as a hard disk, cloud storage and the like, and an interface for outputting the data is provided for other systems or personnel to review and use. Meanwhile, visual products such as reports or charts can be generated, and the visual products are convenient for non-professional staff to understand and use.
The invention provides a computer readable storage medium, on which a computer program is stored, the computer program executing a control point layout module, a sub-region dividing module, a data processing and analyzing module, a storage and output module, a flight environment evaluation module, a navigation measurement execution module, an environment data monitoring module and a knowledge graph construction module when running. The storage medium can be a computer hard disk, an optical disk, a U disk and the like, and is convenient for data transmission and storage.
The invention determines the position of a ground control point through a convolutional neural network and reinforcement learning; and an optimal path for ground control point layout is generated through an optimization algorithm, so that mapping efficiency is improved. The change of the earth surface type, the space distribution condition and the change quantity can be rapidly obtained through the periodical aerial survey and image segmentation technology, and the mapping efficiency is further improved. By arranging the air environment data sensor, the air environment data of each sub-area is monitored in real time, and an important basis is provided for adjustment of a flight plan. By combining unmanned aerial vehicle model, camera parameters, air environment data and weather report data to construct a knowledge graph, and deducing whether a flight environment can execute a flight plan or not through the knowledge graph, flight mapping under proper conditions is ensured, and accuracy of mapping results is improved. Through the data processing and analyzing module, the change trend and the space distribution condition of the earth surface type are revealed, decision basis is provided for homeland planning and resource management, understanding and use by non-professional staff are facilitated, and decision support capability is improved.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (12)
1. A homeland mapping method based on a dynamic remote sensing technology is characterized by comprising the following steps:
S1, dividing a region to be painted into different sub-regions;
s2, determining the position of a ground control point by using a convolutional neural network and reinforcement learning, performing semantic segmentation on the aerial image by using the convolutional neural network CNN, dividing the image into different categories, and extracting the ground feature information of the image; training an agent using the RL, the object of the agent being to select a ground control point on the image; the intelligent agent is used for selecting ground control points on satellite images, outputting the positions and the numbers of the ground control points, and completing an automatic ground control point selecting task; generating an optimal path laid by the ground control points of each sub-area by using an optimization algorithm; the optimal path determining method for the ground control point layout of each sub-area comprises the following steps:
s208, regarding each ground control point as a node, adding a directed edge between two ground control points, and giving a weight to the edge, wherein the weight represents the distance from one point to the other point;
s209, setting the distance of each node to infinity, and setting the distance of the starting point to 0;
s210, using a minimum heap method, taking node distance as a sequencing basis, and only including a starting point at the beginning;
S211, taking out the node with the smallest currently known distance from the priority queue in each round, checking all adjacent nodes of the node, and updating the distance value; if the distance value is updated, marking the node as accessed and adding the node to a priority queue;
S212, stopping the algorithm when the priority queue is empty or the currently fetched node is a target node;
S213, by backtracking the precursor node of each node, the shortest path from the starting point to the end point can be obtained;
s3, setting an air environment data sensor in each sub-area, and detecting real-time environment data of each sub-area;
s4, constructing a knowledge graph according to the callable unmanned aerial vehicle model and camera parameters and combining air environment data and weather report data;
S5, using the knowledge graph to infer whether the flight environment can execute the flight plan, inquiring an evaluation result of the flight environment in the knowledge graph according to the flight plan input by a user, judging whether each sub-area covered by the route can execute the flight plan, and inquiring a matching relationship between the unmanned plane model and the camera parameters in the knowledge graph if the sub-area covered by the route can execute the flight plan; determining unmanned plane models and camera parameters which can be used by a flight environment;
S6, carrying out regular flight mapping on the area to be mapped, and comparing flight mapping results in different periods to determine the change condition.
2. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 1, wherein: in step S2, the method for determining the position of the ground control point includes the steps of:
S201, preparing a large number of marked aviation image data sets, wherein each pixel is marked with a corresponding ground object type label, and the aviation image data sets are divided into a training set and a testing set; collecting satellite images of regions needing territory mapping;
s202, inputting aviation images of a training set into a DeepLabv & lt3+ & gt model constructed by using Pytorch frames for training, and updating model weights through a back propagation algorithm to enable the model to accurately predict the ground object category to which each pixel belongs;
S203, after training convergence, evaluating the performance of the model by using a test set, and carrying out pixel-level semantic segmentation prediction on satellite images of areas needing to be subjected to homeland mapping to obtain DeepLabv3+ segmentation graphs;
S204, flattening the DeepLabv3+ segmentation map, and converting the three-dimensional tensor of H multiplied by W multiplied by C into a one-dimensional vector form with the length of HWC to be used as the state input of the intelligent agent; wherein H is the image height, i.e. the pixel row number, W is the image width, i.e. the pixel column number, C is the channel number, HWC is the length of the one-dimensional vector, hwc=h×w×c, and the meaning of "length HWC" means that the number of elements of this flattened one-dimensional vector is equal to the product of the height, width and channel number of the original three-dimensional tensor;
s205, defining an agent, wherein the output format of the agent is an actual coordinate set of a possible control point;
S206, designing a reward function: after each step of selection, calculating the scores of the possible control points selected at the time in terms of geometric intensity, definition and elevation, and giving a reward value;
s207, selecting an optimal control point by using the deep Q learning DQN through continuous trial and error learning as a ground control point selection strategy.
3. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 2, wherein: determining the position of a ground control point through a convolutional neural network and reinforcement learning;
The reward function of the agent is designed based on the geometric intensity, target definition and elevation relief factors of the ground control points to encourage the agent to select the appropriate ground control points.
4. A homeland mapping method based on dynamic remote sensing technology as claimed in claim 3, wherein:
the reward function determines the selection of ground control points through three factors of geometric intensity, target definition and elevation fluctuation, and introduces a weight vector W= [ W G,wC,wE ] which respectively corresponds to the importance of the three factors;
The bonus function R (x, y) is: r (x, y) =w GG(x,y)+wCC(x,y)+wE E (x, y), where G (x, y) represents the geometric intensity of the ground control point selected at the location (x, y), measures the distribution uniformity of the newly selected point relative to the selected ground control point, C (x, y) represents the target definition at the location (x, y), determined according to the semantic segmentation result, and E (x, y) represents the elevation fluctuation condition at the location (x, y);
G (x, y) =1/d avg(x,y),davg (x, y) is the average distance of location (x, y) to the selected ground control point;
C (x, y) =f clarity(segmentation_map[y,x]),fclarity is a mapping function that converts the semantic segmentation labels into sharpness scores;
E (x, y) = |elevation_data [ x, y ] - μ| if elevation_data is present, μ is the average elevation of the selected ground control point, and the absolute difference may reflect the degree of elevation relief.
5. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 1, wherein: in step S2, when the ground control points are distributed, the corrosion-resistant materials are used for distribution, so that the regular mapping and long-term use are facilitated.
6. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 1, wherein: in step S4, the knowledge graph construction method includes the following steps:
s41, acquiring real-time environment data, weather report data of each sub-area, callable unmanned aerial vehicle model and camera parameter information, storing the information as three data tables, performing data preprocessing, and adding a main key for each data table;
S42, identifying the entities in the knowledge graph from the three data tables, adding a label for each entity to represent the entity type, and adding a name for each entity;
s43, extracting the relation among the entities from the three data tables, assigning a type for each relation, representing the nature of the relation, assigning a direction for each relation, and representing the direction of the relation;
s44, constructing an ontology according to the types of the entities and the relations, defining the attribute of each entity type, defining the attribute of each relation type, and defining the constraint condition of each entity type and each relation type;
S45, storing the entity, the relation and the attribute in a graph database to form a knowledge base of a graph structure;
S46, utilizing the ontology and logic rules to infer data in the knowledge base, supplementing missing information, finding new knowledge, improving the quality and the integrity of the knowledge, and inferring an evaluation result of the flight environment of each sub-area according to the influence relationship of air environment data and weather report data on the flight environment;
S47, inquiring an evaluation result of the flight environment in the knowledge graph according to the flight plan input by the user, and judging whether the flight plan can be executed or not; if the unmanned aerial vehicle can fly, inquiring the matching relation between the unmanned aerial vehicle model and the camera parameters in the knowledge graph, selecting the proper unmanned aerial vehicle model and the proper camera parameters, making a flight plan, and feeding back the information to the user.
7. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 6, wherein: in step S43, the identifier, the tag, the name, the attribute and the constraint condition information are added to each node and each side, and an index is added to the knowledge base, so that the data can be conveniently and rapidly queried and retrieved.
8. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 6, wherein: in step S46, according to the matching relationship between the unmanned aerial vehicle model and the camera parameters, the camera parameters that can be used by each unmanned aerial vehicle model are deduced, and these information are stored as relationships and attributes in the knowledge base.
9. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 1, wherein: real-time environmental data includes wind speed, wind direction, humidity, particulate matter concentration, illumination intensity and luminousness, unmanned aerial vehicle model information includes model, weight, duration, load and flight speed, camera parameter information includes model, resolution, focus and aperture of camera, and weather report data includes weather, atmospheric pressure, visibility and precipitation.
10. The homeland mapping method based on dynamic remote sensing technology as set forth in claim 1, wherein: in step S6, the change of the earth surface type, the space distribution condition and the change amount are judged by comparing the classes of the flight mapping image pixels in different periods through image segmentation; and (3) determining the change of the terrain by periodically measuring the position information of the ground control point.
11. A homeland mapping system based on dynamic remote sensing technology, comprising: the system comprises a control point layout module, a subarea division module, a data processing and analyzing module, a storage and output module, a flight environment assessment module, a aerial survey execution module, an environment data monitoring module and a knowledge graph construction module; the method is characterized in that:
Control point layout module: determining the position of a ground control point by using a convolutional neural network and reinforcement learning, and generating an optimal path laid by the ground control point of each sub-area by using an optimization algorithm;
The subarea dividing module: dividing the region to be painted into different subareas according to the set flight altitude, flight speed and flight plan content of the route;
The environmental data monitoring module: an air environment data sensor is arranged in each sub-area and used for detecting air environment data of each sub-area in real time, wherein the air environment data comprise temperature, humidity, wind speed, wind direction and atmospheric pollutant concentration;
knowledge graph construction module: combining the callable unmanned aerial vehicle model, camera parameters and air environment data and weather report data acquired in real time to construct a knowledge graph;
Flight environment assessment module: utilizing the knowledge graph to infer whether the flight environment can execute a flight plan, and determining the usable unmanned aerial vehicle model and camera parameters;
The aerial survey execution module: performing regular aerial survey on the region to be mapped, and judging the change of the earth surface type, the spatial distribution condition and the change quantity by using an image segmentation technology;
And the data processing and analyzing module is used for: collecting and processing data generated by each module, and revealing the variation trend and the spatial distribution condition of the earth surface type by analyzing and comparing mapping results of different time points;
and a storage and output module: the collected data and the processed data are stored in a suitable storage medium.
12. A computer-readable storage medium, characterized by: on which a computer program is stored which, when run, implements the method of any one of claims 1 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410294927.4A CN117907242A (en) | 2024-03-15 | 2024-03-15 | Homeland mapping method, system and storage medium based on dynamic remote sensing technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410294927.4A CN117907242A (en) | 2024-03-15 | 2024-03-15 | Homeland mapping method, system and storage medium based on dynamic remote sensing technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117907242A true CN117907242A (en) | 2024-04-19 |
Family
ID=90692632
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410294927.4A Pending CN117907242A (en) | 2024-03-15 | 2024-03-15 | Homeland mapping method, system and storage medium based on dynamic remote sensing technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117907242A (en) |
Citations (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180372493A1 (en) * | 2017-06-27 | 2018-12-27 | Infatics, Inc. (DBA DroneDeploy) | Method and system for aerial image processing |
| CN109815916A (en) * | 2019-01-28 | 2019-05-28 | 成都蝉远科技有限公司 | A kind of recognition methods of vegetation planting area and system based on convolutional neural networks algorithm |
| KR102003187B1 (en) * | 2019-06-13 | 2019-07-23 | 주식회사 디지털커브 | Method and apparatus for modeling resultant image to which each of ground control points including location information of survey site is matched using images captured by unmanned air vehicle |
| CN110222127A (en) * | 2019-06-06 | 2019-09-10 | 中国电子科技集团公司第二十八研究所 | The converging information method, apparatus and equipment of knowledge based map |
| US20210221506A1 (en) * | 2018-10-02 | 2021-07-22 | Robert S. Phelan | Unmanned aerial vehicle system and methods |
| CN113626616A (en) * | 2021-08-25 | 2021-11-09 | 中国电子科技集团公司第三十六研究所 | Aircraft safety early warning method, device and system |
| US20210407126A1 (en) * | 2017-05-04 | 2021-12-30 | Skydio, Inc. | Ground control point center determination |
| WO2022003740A1 (en) * | 2020-07-02 | 2022-01-06 | Alma Mater Studiorum - Universita' Di Bologna | Method for determining the confidence of a disparity map through a self-adaptive learning of a neural network, and sensor system thereof |
| CN114022784A (en) * | 2021-11-09 | 2022-02-08 | 中国人民解放军61646部队 | Method and device for screening landmark control points |
| CN114120141A (en) * | 2021-11-23 | 2022-03-01 | 深圳航天智慧城市系统技术研究院有限公司 | All-weather remote sensing monitoring automatic analysis method and system thereof |
| CN114518104A (en) * | 2022-03-14 | 2022-05-20 | 山东三津房地产评估有限公司 | Territorial surveying and mapping method, system and storage medium based on dynamic remote sensing monitoring technology |
| CN115034696A (en) * | 2022-08-12 | 2022-09-09 | 北京迈平测绘技术开发有限公司 | Geological mapping cooperative operation method and system |
| CN115147730A (en) * | 2022-07-22 | 2022-10-04 | 厦门天卫科技有限公司 | Remote sensing classification method combining full convolution neural network and ensemble learning |
| KR20220169342A (en) * | 2021-06-18 | 2022-12-27 | 이재영 | Drone used 3d mapping method |
| CN116182805A (en) * | 2023-02-17 | 2023-05-30 | 山东盛优佳生物防控有限责任公司 | Homeland mapping method based on remote sensing image |
| CN116539004A (en) * | 2023-05-04 | 2023-08-04 | 中国人民解放军国防科技大学 | Communication line engineering investigation design method and system adopting unmanned aerial vehicle mapping |
| CN116958825A (en) * | 2023-08-28 | 2023-10-27 | 中国公路工程咨询集团有限公司 | Mobile remote sensing image acquisition method and highway maintenance monitoring method |
| WO2023223284A1 (en) * | 2022-05-19 | 2023-11-23 | Ai Clearing Inc. | System and method for triggering data transfer using progress tracking |
| US20230419501A1 (en) * | 2020-11-16 | 2023-12-28 | Ai Clearing Inc. | Image analysis for aerial images |
| CN117408536A (en) * | 2023-12-15 | 2024-01-16 | 山东省物化探勘查院 | AI-based underwater mapping real-time analysis system |
| CN117493818A (en) * | 2024-01-03 | 2024-02-02 | 山东亿华天产业发展集团有限公司 | Homeland mapping method, system and storage medium based on dynamic remote sensing technology |
| CN117496073A (en) * | 2023-12-29 | 2024-02-02 | 山东省国土测绘院 | Method and system for constructing multi-time-phase live-action three-dimensional model |
| CN117521932A (en) * | 2023-11-22 | 2024-02-06 | 国网福建省电力有限公司电力科学研究院 | Unmanned aerial vehicle inspection management system based on meshing division |
| CN117570881A (en) * | 2024-01-17 | 2024-02-20 | 贵州省第一测绘院(贵州省北斗导航位置服务中心) | Land area measurement system and method for investigation of homeland resources |
| US20240070808A1 (en) * | 2021-01-15 | 2024-02-29 | Medici Land Governance | Enhancement of coarse resolution imagery using super-resolution and convolution neural networks to be used for fit-for-purpose systematic land titling and land valuation |
-
2024
- 2024-03-15 CN CN202410294927.4A patent/CN117907242A/en active Pending
Patent Citations (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210407126A1 (en) * | 2017-05-04 | 2021-12-30 | Skydio, Inc. | Ground control point center determination |
| US20180372493A1 (en) * | 2017-06-27 | 2018-12-27 | Infatics, Inc. (DBA DroneDeploy) | Method and system for aerial image processing |
| US20210221506A1 (en) * | 2018-10-02 | 2021-07-22 | Robert S. Phelan | Unmanned aerial vehicle system and methods |
| CN109815916A (en) * | 2019-01-28 | 2019-05-28 | 成都蝉远科技有限公司 | A kind of recognition methods of vegetation planting area and system based on convolutional neural networks algorithm |
| CN110222127A (en) * | 2019-06-06 | 2019-09-10 | 中国电子科技集团公司第二十八研究所 | The converging information method, apparatus and equipment of knowledge based map |
| KR102003187B1 (en) * | 2019-06-13 | 2019-07-23 | 주식회사 디지털커브 | Method and apparatus for modeling resultant image to which each of ground control points including location information of survey site is matched using images captured by unmanned air vehicle |
| WO2022003740A1 (en) * | 2020-07-02 | 2022-01-06 | Alma Mater Studiorum - Universita' Di Bologna | Method for determining the confidence of a disparity map through a self-adaptive learning of a neural network, and sensor system thereof |
| US20230419501A1 (en) * | 2020-11-16 | 2023-12-28 | Ai Clearing Inc. | Image analysis for aerial images |
| US20240070808A1 (en) * | 2021-01-15 | 2024-02-29 | Medici Land Governance | Enhancement of coarse resolution imagery using super-resolution and convolution neural networks to be used for fit-for-purpose systematic land titling and land valuation |
| KR20220169342A (en) * | 2021-06-18 | 2022-12-27 | 이재영 | Drone used 3d mapping method |
| CN113626616A (en) * | 2021-08-25 | 2021-11-09 | 中国电子科技集团公司第三十六研究所 | Aircraft safety early warning method, device and system |
| CN114022784A (en) * | 2021-11-09 | 2022-02-08 | 中国人民解放军61646部队 | Method and device for screening landmark control points |
| CN114120141A (en) * | 2021-11-23 | 2022-03-01 | 深圳航天智慧城市系统技术研究院有限公司 | All-weather remote sensing monitoring automatic analysis method and system thereof |
| CN114518104A (en) * | 2022-03-14 | 2022-05-20 | 山东三津房地产评估有限公司 | Territorial surveying and mapping method, system and storage medium based on dynamic remote sensing monitoring technology |
| WO2023223284A1 (en) * | 2022-05-19 | 2023-11-23 | Ai Clearing Inc. | System and method for triggering data transfer using progress tracking |
| CN115147730A (en) * | 2022-07-22 | 2022-10-04 | 厦门天卫科技有限公司 | Remote sensing classification method combining full convolution neural network and ensemble learning |
| CN115034696A (en) * | 2022-08-12 | 2022-09-09 | 北京迈平测绘技术开发有限公司 | Geological mapping cooperative operation method and system |
| CN116182805A (en) * | 2023-02-17 | 2023-05-30 | 山东盛优佳生物防控有限责任公司 | Homeland mapping method based on remote sensing image |
| CN116539004A (en) * | 2023-05-04 | 2023-08-04 | 中国人民解放军国防科技大学 | Communication line engineering investigation design method and system adopting unmanned aerial vehicle mapping |
| CN116958825A (en) * | 2023-08-28 | 2023-10-27 | 中国公路工程咨询集团有限公司 | Mobile remote sensing image acquisition method and highway maintenance monitoring method |
| CN117521932A (en) * | 2023-11-22 | 2024-02-06 | 国网福建省电力有限公司电力科学研究院 | Unmanned aerial vehicle inspection management system based on meshing division |
| CN117408536A (en) * | 2023-12-15 | 2024-01-16 | 山东省物化探勘查院 | AI-based underwater mapping real-time analysis system |
| CN117496073A (en) * | 2023-12-29 | 2024-02-02 | 山东省国土测绘院 | Method and system for constructing multi-time-phase live-action three-dimensional model |
| CN117493818A (en) * | 2024-01-03 | 2024-02-02 | 山东亿华天产业发展集团有限公司 | Homeland mapping method, system and storage medium based on dynamic remote sensing technology |
| CN117570881A (en) * | 2024-01-17 | 2024-02-20 | 贵州省第一测绘院(贵州省北斗导航位置服务中心) | Land area measurement system and method for investigation of homeland resources |
Non-Patent Citations (7)
| Title |
|---|
| HUANG, SHENJIN 等: "Recognizing Zucchinis Intercropped with Sunflowers in UAV Visible Images Using an Improved Method Based on OCRNet", 《REMOTE SENSING》, vol. 13, no. 14, 31 July 2021 (2021-07-31), pages 2706 * |
| ZHAO, LIXING 等: "A CNN-Based Layer-Adaptive GCPs Extraction Method for TIR Remote Sensing Images", 《REMOTE SENSING》, vol. 15, no. 10, 31 May 2023 (2023-05-31), pages 2628 * |
| 中国科学技术协会: "《2018-2019年学科发展报告综合卷》", vol. 1, 30 November 2021, 中国科学技术出版社, pages: 250 - 254 * |
| 孙钰 等: "基于深度学习的大棚及地膜农田无人机航拍监测方法", 《农业机械学报》, vol. 49, no. 02, 25 February 2018 (2018-02-25), pages 133 - 140 * |
| 张振超: "基于机载激光扫描与摄影测量数据的变化检测技术研究", 《中国博士学位论文全文数据库 基础科学辑》, no. 3, 15 March 2022 (2022-03-15), pages 008 - 41 * |
| 石川: "遥感测绘技术在监测土壤污染中的应用研究", 《中国新技术新产品》, vol. 6, 25 June 2023 (2023-06-25), pages 130 - 132 * |
| 贺书 等: "贵州农业产业遥感调查监测与3S技术在茶产业中的应用", 《农技服务》, vol. 38, no. 5, 20 May 2021 (2021-05-20), pages 73 - 75 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10168141B2 (en) | Method for identifying air pollution sources based on aerosol retrieval and glowworm swarm algorithm | |
| CN110490415B (en) | Building progress assessment method for visual coordination of multiple unmanned aerial vehicles | |
| Guo et al. | Ecomark 2.0: empowering eco-routing with vehicular environmental models and actual vehicle fuel consumption data | |
| CN112561191B (en) | Prediction model training method, prediction device, prediction apparatus, prediction program, and program | |
| CN112308292B (en) | Method for drawing fire risk grade distribution map | |
| CN109711751A (en) | A kind of resosurces environment loading capacity appraisal procedure based on space-time data convergence analysis | |
| CN106485360A (en) | Segmental society's prediction of economic indexes method and system based on overall noctilucence remote sensing | |
| CN116595121B (en) | Data display monitoring system based on remote sensing technology | |
| CN116018602A (en) | Vegetation management system and vegetation management method | |
| CN113360587A (en) | Land surveying and mapping equipment and method based on GIS technology | |
| CN111626103A (en) | Urbanization gradient measurement method based on night light remote sensing image | |
| CN116415499B (en) | Community comfort simulation prediction method | |
| CN117907242A (en) | Homeland mapping method, system and storage medium based on dynamic remote sensing technology | |
| CN116324523A (en) | Turbulence prediction system and turbulence prediction method | |
| KR20050063615A (en) | Method for providing surface roughness in geographic information system | |
| Bastani et al. | SkyQuery: an aerial drone video sensing platform | |
| US11842537B1 (en) | Devices, methods, and graphical user interfaces for analyzing, labeling, and managing land in a geospatial platform | |
| KR102470201B1 (en) | system and method for providing risk of building wind | |
| CN117270913B (en) | Map updating method, device, electronic equipment and storage medium | |
| RU2796370C1 (en) | System and method for dynamic visualization of environmental pollution | |
| CN117458450B (en) | Power data energy consumption prediction analysis method and system | |
| Verstraete | Artificial Intelligent Methods for Handling Spatial Data: Fuzzy Rulebase Systems and Gridded Data Problems | |
| Hemmes | Classification of large scale outdoor point clouds using convolutional neural networks | |
| CN118012977B (en) | AI and GIS fusion-based two-dimensional multi-mode data processing method | |
| CN117708551B (en) | Flood disaster influence assessment method and system based on double-precision GDP data distribution |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |