Wetland observation method, device, equipment and medium based on band simulation

Abstract

The embodiment of the present application discloses a wetland observation method, device, equipment and medium based on band simulation, which relates to the field of satellite remote sensing technology. The method includes: obtaining satellite multispectral data, obtaining drone hyperspectral data, inputting satellite multispectral data and drone hyperspectral data into a trained band simulation model, so that the trained band simulation model simulates and obtains simulated hyperspectral data of the target area; analyzing the simulated hyperspectral data based on characteristic bands by a trained land object classification model, and outputting land object distribution data of the target area. By selecting several sample areas within the target area, the present application only needs to collect hyperspectral data within the sample area to fit the simulated hyperspectral data of the entire target area, and can efficiently output hyperspectral data for different land objects in different regions, providing more comprehensive information for ecological monitoring.

Description

Technical Field

[0001] The present application relates to the field of satellite remote sensing technology, and in particular to a wetland observation method, device, equipment and medium based on band simulation. Background Art

[0002] Wetlands are important and unique ecosystems on Earth, with many functions, including water conservation, water purification, climate regulation, carbon fixation and oxygen release, and biodiversity maintenance. Due to the impact of climate change and human activities, the area of ​​wetlands is decreasing year by year, and the biodiversity of wetland ecosystems is continuously decreasing. Therefore, it is very important to strengthen the protection of wetlands and wetland ecosystems, and wetland monitoring is of great significance to the protection of wetland ecosystems.

[0003] With the rapid development of remote sensing technology, the application fields of remote sensing are becoming more and more extensive. Satellite remote sensing technology is widely used around the world. Various satellite sensors can provide multi-spectral, high-resolution remote sensing data. The improvement of resolution and spectral range enables satellite remote sensing to capture surface features more accurately, thus providing multi-spectral data over a large range. However, although the multi-spectral data provided by satellite sensors can cover a large range, the acquired bands are relatively small and the spectral resolution is limited, which makes it difficult to meet the demand for subtle spectral features.

[0004] In addition, UAV remote sensing takes advantage of the high flexibility and strong real-time performance of UAVs, and can obtain detailed information of the target area with a high spatial resolution. However, the area that a UAV can collect at one time is small, and multiple flights are required to obtain a large range of hyperspectral data, which will bring additional time and labor costs. In the process of collecting a large range of hyperspectral data, it is also easy to miss some areas, or there is overlap in the hyperspectral data collected twice, resulting in low efficiency in collecting hyperspectral data.

[0005] Therefore, even if the relevant technology combines the multispectral data of satellite remote sensing and the hyperspectral data of UAV remote sensing, the method of gradually collecting hyperspectral data only through UAVs is inefficient, which makes it difficult to obtain hyperspectral data, and further leads to limitations in collecting spectral data in certain large-area scenes. Summary of the invention

[0006] The present application provides a wetland observation method, device, equipment and medium based on band simulation, aiming to solve the problem in the prior art that it is difficult to combine multispectral data from satellite remote sensing and hyperspectral data from drone remote sensing due to the limited efficiency of hyperspectral data collection. The technical solution is as follows:

[0007] In a first aspect, an embodiment of the present application provides a wetland observation method based on band simulation, comprising:

[0008] Obtain satellite multispectral data of the target area and obtain drone hyperspectral data of a preset number of sample areas;

[0009] The satellite multispectral data and the UAV hyperspectral data are input into a trained band simulation model, so that the trained band simulation model simulates and obtains the simulated hyperspectral data of the target area according to the mapping relationship between the UAV hyperspectral data in each sample area and the satellite multispectral data in the corresponding area;

[0010] Inputting the simulated hyperspectral data into a trained land object classification model, so that the trained land object classification model analyzes the simulated hyperspectral data based on characteristic bands and outputs land object distribution data of the target area;

[0011] Wherein, each of the sample areas does not overlap with each other and is located within the range of the target area.

[0012] In an optional solution of the first aspect, after acquiring satellite multispectral data of the target area and acquiring drone hyperspectral data of a preset number of sample areas, the method further includes:

[0013] Preprocessing the satellite multispectral data and the drone hyperspectral data, and aligning the satellite multispectral data with the drone hyperspectral data of the preset number of sample areas based on geographic location information;

[0014] Performing radiation correction and atmospheric correction on the aligned data, and outputting the corrected data;

[0015] The step of inputting the satellite multispectral data and the drone hyperspectral data into a trained band simulation model comprises:

[0016] The corrected data is input into the trained band simulation model.

[0017] In an optional solution of the first aspect, the trained band simulation model is trained according to the following steps, including:

[0018] The drone hyperspectral data of a preset area and the satellite multispectral data of the preset area are used as a sample set;

[0019] The sample set is input into the band simulation model, each band of the satellite multispectral data in the preset area is used as an independent variable, and each band of the drone hyperspectral data in the preset area is used as a dependent variable, so that the band simulation model learns the mapping relationship between the reflectivity of the drone hyperspectral data in the preset area and the reflectivity of the satellite multispectral data;

[0020] Output the trained band simulation model.

[0021] In an optional solution of the first aspect, before acquiring the drone hyperspectral data of a preset number of sample areas, the method further includes:

[0022] estimating the types of objects in the target area and the estimated distribution data of various objects in the target area;

[0023] A preset number of sample areas are selected within the target area based on the estimated distribution data; wherein each of the sample areas contains at least one of the land features, and the set of land feature types contained in all the sample areas includes all land feature types in the target area.

[0024] In an optional solution of the first aspect, after acquiring the satellite multispectral data of the target area and before acquiring the drone hyperspectral data of a preset number of sample areas, the method further includes:

[0025] Determine the spectral reflectance distribution in the target area according to the satellite multispectral data;

[0026] The area with the lowest spectral reflectance and the area with the highest spectral reflectance in the target area are determined according to the spectral reflectance distribution, and the corresponding area is determined as the sample area.

[0027] In an optional solution of the first aspect, before inputting the simulated hyperspectral data into the trained land object classification model, the method further includes:

[0028] Obtain the reflectivity of each ground feature in each band;

[0029] Obtain the reflectivity difference of various objects in the same band;

[0030] The characteristic band for identifying the type of the land feature is determined based on the reflectivity and the reflectivity difference.

[0031] In an optional solution of the first aspect, the step of acquiring satellite multispectral data of the target area and acquiring drone hyperspectral data of a preset number of sample areas is performed at intervals of a preset time period;

[0032] After outputting the object distribution data of the target area, the method further includes:

[0033] The change trend of the ground object distribution data of the target area is calculated based on the ground object distribution data of the target area in each of the preset time periods.

[0034] In a second aspect, the embodiment of the present application further provides a wetland observation device based on band simulation, comprising:

[0035] The spectral data acquisition module is used to acquire satellite multispectral data of the target area and obtain UAV hyperspectral data of a preset number of sample areas;

[0036] A band simulation module is used to input the satellite multispectral data and the UAV hyperspectral data into a trained band simulation model, so that the trained band simulation model simulates and obtains the simulated hyperspectral data of the target area according to the mapping relationship between the UAV hyperspectral data in each sample area and the satellite multispectral data in the corresponding area;

[0037] A ground object classification module, used for inputting the simulated hyperspectral data into a trained ground object classification model, so that the trained ground object classification model analyzes the simulated hyperspectral data based on characteristic bands and outputs ground object distribution data of the target area;

[0038] Wherein, each of the sample areas does not overlap with each other and is located within the range of the target area.

[0039] In a third aspect, an embodiment of the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method provided in the first aspect of the embodiment of the present application or any one of the implementation methods of the first aspect is implemented.

[0040] In a fourth aspect, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method provided by the first aspect of the embodiment of the present application or any one of the implementations of the first aspect.

[0041] The beneficial effects brought about by the technical solutions provided by some embodiments of the present application include at least:

[0042] The wetland observation method, device, equipment and medium based on band simulation provided by the embodiment of the present application, simulates the simulated hyperspectral data of the corresponding area according to the mapping relationship between the drone hyperspectral data and the satellite multispectral data in the corresponding area through the trained band simulation model, and integrates the high coverage of satellite multispectral data and the flexibility of drone hyperspectral technology, thereby making up for the defects of multispectral data with wide coverage but few bands and hyperspectral data with wide bands but small range. In this way, the demand for hyperspectral information in large-scale scenes can be met, and the defects and limitations of related technologies that cannot efficiently obtain hyperspectral data can be overcome. The present application also selects several sample areas in the target area, and only needs to collect the hyperspectral data in the sample area to fit the simulated hyperspectral data of the entire target area. Hyperspectral data can be efficiently output for different objects in different regions. By selecting the sample area as a typical area for identifying objects, the efficiency and accuracy of simulated hyperspectral data can be targetedly improved, providing more comprehensive information for ecological monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the present application or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 is a schematic diagram of the architecture of a wetland observation system based on band simulation according to an embodiment of the present application;

[0045] Figure 2 It is a flow chart of a wetland observation method based on band simulation according to an embodiment of the present application;

[0046] Figure 3 is a schematic diagram of a spectral curve of a typical wetland feature provided in an embodiment of the present application;

[0047] Figure 4 It is a structural schematic diagram of a wetland observation device based on band simulation provided in an embodiment of the present application;

[0048] Figure 5 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] The terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or modules is not limited to the listed steps or modules, but may optionally include steps or modules that are not listed, or may optionally include other steps or modules that are inherent to these processes, methods, products or devices.

[0051] It should be noted that the terms "first\second" involved in the present application are only used to distinguish similar objects, and do not represent a specific order for the objects. It is understandable that "first\second" can be interchanged with a specific order or sequence where permitted. It should be understood that the objects distinguished by "first\second" can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those described or illustrated herein.

[0052] Please refer to the following Figure 1 , which is a schematic diagram of the architecture of a wetland observation system based on band simulation provided by an exemplary embodiment of the present application. Figure 1 As shown, the wetland observation and computing system based on band simulation includes a data acquisition unit 110 and a computing unit 120 .

[0053] The data acquisition unit 110 may include a device for collecting hyperspectral data and multispectral data, and may specifically include one or more drones for collecting hyperspectral data, and may also include one or more satellites for collecting multispectral data, and may also include a terminal device for receiving data transmitted back by drones and / or satellites, which is not limited in the present embodiment of the application. The user may use the data acquisition unit 110 to collect satellite multispectral data of the target area, and collect drone hyperspectral data of a preset number of sample areas, and send the collected satellite multispectral data and drone hyperspectral data to the calculation unit 120.

[0054] The computing unit 120 may be a computing terminal that can provide a variety of wetland observation methods based on band simulation. The computing unit 120 may be a terminal device such as a personal computer, workstation, mobile terminal, server, virtual server, cloud server, etc. that has data computing capabilities. The computing unit 120 is equipped with a trained band simulation model and a ground object classification model. The computing unit 120 may receive satellite multispectral data and drone hyperspectral data of the target area collected by the data acquisition unit 110 through the network, and obtain simulated hyperspectral data by simulating the mapping relationship between the drone hyperspectral data in each sample area and the satellite multispectral data in the corresponding area through the trained band simulation model. The computing unit 120 may also analyze the simulated hyperspectral data based on the characteristic bands through the trained ground object classification model, and output the ground object distribution data of the target area to the data acquisition unit 110.

[0055] The calculation process of the wetland observation based on band simulation is not limited to the calculation unit 120, but can also be performed separately by the data acquisition unit 110. The present application does not specifically limit this, and the following embodiments are all described by taking the data acquisition unit 110 performing the wetland observation method based on band simulation as an example.

[0056] The network may be a medium providing a communication link between the computing unit 120 and the data acquisition unit 110, or may be the Internet including network devices and transmission media, but is not limited thereto. The transmission media may be a wired link (such as but not limited to coaxial cable, optical fiber, and digital subscriber line (DSL)) or a wireless link (such as but not limited to wireless fidelity (WIFI), Bluetooth, and mobile device network).

[0057] Understandably, Figure 1 The number of computing units 120 and data acquisition units 110 in the wetland observation system based on band simulation is shown as an example only. In a specific implementation, the system may include any number of computing units 120 and data acquisition units 110. This specification embodiment does not specifically limit this. For example, but not limited to, the data acquisition unit 110 may be a data acquisition unit cluster composed of multiple data acquisition units, and the computing unit 120 may be a computing unit cluster composed of multiple computing units.

[0058] The present application is described in detail below with reference to specific embodiments.

[0059] Next, combine Figure 1 , taking the wetland observation method based on band simulation performed by the data acquisition unit as an example, the wetland observation method based on band simulation provided by the embodiment of the present application is introduced. Figure 2 ,like Figure 2 As shown, the method comprises the following steps:

[0060] S201, obtaining satellite multispectral data of a target area, and obtaining drone hyperspectral data of a preset number of sample areas.

[0061] Specifically, in general, satellite multispectral data includes fewer bands, while drone hyperspectral data can include up to 300 bands, far exceeding traditional multispectral sensors.

[0062] It should be noted that there is no overlapping area between each sample area, and all sample areas are located within the range of the target area.

[0063] Specifically, an area in the wetland that needs to be observed can be selected as a target area to be observed, and multispectral data in the target area can be acquired through a satellite.

[0064] Specifically, drone hyperspectral data can be obtained by operating a drone to measure in the target area. The hyperspectral data obtained by the drone each time can cover an area greater than or equal to the sample area. The part larger than the sample area can be cropped to obtain drone hyperspectral data of the sample area.

[0065] It should be noted that the sample area should be an area that can characterize the characteristics of the landform type in the target area.

[0066] In some embodiments, the types of land features in the target area and the distribution data can be estimated by field measurements, consulting wetland information provided by the forestry department, or observing satellite images.

[0067] It should be noted that the land features include the surface types in the wetland, such as water bodies, soil, concrete and other terrains, and also include vegetation composed of various types of plants, such as lotus, calamus and other plant species.

[0068] For example, for a certain area A in a wetland, it can be estimated that there are at least water bodies, reeds, lotus and other features in area A. It can also be estimated that there are areas with more water bodies, areas with more reed populations and areas with more lotus distribution, that is, estimated distribution data.

[0069] Specifically, the enrichment area of ​​each type of land feature can be roughly determined by estimating the distribution data, and then the corresponding enrichment area can be determined as a typical area representing the corresponding target area.

[0070] Specifically, each sample area should contain at least one type of feature that the target area has, and the set of feature types of all sample areas should include all feature types of the target area.

[0071] Exemplarily, area A can be used as the target area, and multiple sample areas can be selected. The sample area can be an area containing water, the sample area can be an area containing reeds, the sample area can be an area containing lotus, the sample area can also be an area including two of the features, and the sample area can also be an area including three features in area A.

[0072] It can be understood that the greater the number of sample areas, the larger the range covered by the sample areas, and the easier it is to simulate and obtain high-precision hyperspectral data. The embodiment of the present application does not limit the number of sample areas.

[0073] In some embodiments, the spectral reflectance distribution of the entire area can also be determined through multi-spectral data collected by satellites, and then the areas with the lowest and highest spectral reflectance values ​​can be determined. The areas corresponding to the extreme spectral reflectance values ​​can be used as sample areas, which is beneficial to improving the simulation credibility of the trained band simulation model for the intermediate values.

[0074] It can be understood that both the satellite multispectral data and the UAV hyperspectral data have corresponding geographic coordinates, so it is not difficult to correspond the satellite multispectral data and the UAV hyperspectral data to the actual geographical environment based on the coordinates.

[0075] In some embodiments, after S201, before inputting the satellite multispectral data and the UAV hyperspectral data into the trained band simulation model, the satellite multispectral data and the UAV hyperspectral data may be preprocessed.

[0076] Specifically, preprocessing includes but is not limited to geometric correction, radiation correction, atmospheric correction and the like.

[0077] Specifically, geometric correction can align the satellite multispectral data with the drone hyperspectral data of the preset number of sample areas based on geographic location information, and can be used to correct image geometric deformation caused by relative movement between the sensor and the surface or terrain changes, so that subsequent land object classification and analysis can be performed based on accurate geographic location information.

[0078] Specifically, radiation correction and atmospheric correction may be performed on the aligned and geometrically corrected data, and the corrected data may be output.

[0079] Among them, radiation correction can be used to eliminate the differences in radiation measurements caused by factors such as atmospheric scattering and sensor response in images collected by satellites and drones. Through this correction step, the true radiation brightness value can be obtained, making the remote sensing data of satellites and drones more objective and can be calculated based on the true radiation brightness value. Radiation correction helps to eliminate brightness differences in the collected images and ensure that the spectral information reflecting the surface characteristics is more accurate.

[0080] Among them, atmospheric correction can eliminate the influence of atmospheric media on remote sensing data. Atmospheric correction can reduce the influence of atmospheric scattering and absorption on remote sensing images, further improve the quality and interpretability of images, and play a vital role in accurately extracting surface features, classifying objects, and detecting changes.

[0081] It should be noted that through these correction steps, the quality and consistency of remote sensing data are ensured, providing a reliable data basis for subsequent geographic information extraction, object type identification, and analysis of object distribution data.

[0082] S202, inputting the satellite multispectral data and the UAV hyperspectral data into the trained band simulation model, so that the trained band simulation model simulates the simulated hyperspectral data of the target area according to the mapping relationship between the UAV hyperspectral data in each sample area and the satellite multispectral data in the corresponding area.

[0083] Specifically, during the training process, the band simulation model learns the mapping relationship between the UAV hyperspectral data in the sample area and the satellite multispectral data in the corresponding area. More specifically, it learns the mapping relationship between the hyperspectral data reflectance and the multispectral data reflectance in the same area. Then, based on the learned mapping relationship, the hyperspectral data can be simulated in the area outside the sample area.

[0084] Specifically, the training process of the band simulation model in the embodiment of the present application may include the following steps:

[0085] The drone hyperspectral data of a preset area and the satellite multispectral data of the preset area are used as a sample set.

[0086] The sample set is input into the band simulation model, each band of the satellite multispectral data in the preset area is used as an independent variable, and each band of the UAV hyperspectral data in the preset area is used as a dependent variable, so that the band simulation model learns the mapping relationship between the reflectivity of the UAV hyperspectral data in the preset area and the reflectivity of the satellite multispectral data;

[0087] Output the trained band simulation model.

[0088] Specifically, the preset area is an area where accurate ground object distribution data is known. The ground object type in the preset area and the detailed distribution range of each ground object type can be obtained through hyperspectral data analysis. The embodiment of the present application does not limit the selection method of the preset area.

[0089] Specifically, the sample set includes satellite multispectral data in the area, drone hyperspectral data in the area, and the types of objects in the area and the distribution range of the object types. The object range can be given in the form of coordinates.

[0090] Specifically, during the training process, the model can learn the mapping relationship between the reflectivity of the UAV hyperspectral data and the reflectivity of the satellite multispectral data.

[0091] Specifically, part of the data can be selected as a test set and a validation set in the process of generating a sample set. It is not difficult to verify whether the band simulation model converges and the accuracy of the band simulation model through the test set and the validation set.

[0092] Specifically, the data in the training set can be used to perform convolution operations of different depths on a pixel-by-pixel basis in combination with surrounding pixels. The specific depth of the convolutional network can be debugged according to different task requirements. For example, the number of network layers can be set to 3, with convolution kernels of sizes 3×3×32, 3×3×16, and 3×3×8 respectively, and each convolution layer is interconnected using dense connections.

[0093] Preferably, a densely connected neural network can be selected as the band simulation model in the embodiment of the present application, which can overcome the problems faced by traditional neural networks during training, such as gradient vanishing, sparse parameters, and insufficient feature reuse.

[0094] S203, inputting the simulated hyperspectral data into a trained land object classification model, so that the trained land object classification model analyzes the simulated hyperspectral data based on characteristic bands and outputs land object distribution data of the target area.

[0095] Specifically, after the simulated hyperspectral data is obtained, a plurality of characteristic bands may be selected to identify the simulated hyperspectral data, which may be specifically judged by a land object classification model.

[0096] It should be noted that different objects on the surface have different spectral responses in the visible light, near-infrared, short-wave infrared, thermal infrared and other bands. For example, vegetation has strong reflection characteristics in the near-infrared band and water has absorption characteristics in the visible light and near-infrared bands.

[0097] Based on this, characteristic bands suitable for specific ground object classification can be extracted according to the differences in the spectral characteristics of the ground objects.

[0098] Specifically, the reflectivity of each ground object in each band can be obtained, and the reflectivity difference of multiple ground objects in the same band can be obtained. Based on the reflectivity and the reflectivity difference, the characteristic band used to identify the ground object type can be determined.

[0099] Optionally, the spectral characteristics of each plant and terrain may be directly obtained by consulting data and utilizing existing characteristic bands. The embodiment of the present application does not limit the method for obtaining the characteristic bands.

[0100] For example, Figure 3 The spectral curves of typical wetland features are shown. Typical wetland features include reeds, cattails, lotus and water. By observing the spectral reflectance of these features in different bands, different plants have different chlorophyll absorption and reflection characteristics in the green and red spectral ranges. Therefore, it can be found that these features have obvious differences in the green band, red band and infrared band.

[0101] For example, reeds and cattails have different reflectances in the green band, which makes them easy to distinguish in the corresponding band. Lotus shows a higher reflectance in the infrared band, which is different from other plants. The lowest reflectance of water bodies in the infrared band also provides a basis for distinguishing them from vegetation categories.

[0102] Exemplarily, specific bands suitable for classification in the target area may be determined: 410 nm (blue band), 550 nm (green band), 735 nm (infrared band).

[0103] It should be noted that these specific bands may be used to distinguish vegetation in the green and red bands and to identify water bodies in the infrared band, thereby improving the accuracy of object classification.

[0104] It should be noted that the spectral characteristics of different ground objects can also be identified through the combination of other characteristic bands, and the embodiments of the present application are not limited to this.

[0105] Specifically, the feature classification model can adopt models such as random forest and Unet. For example, taking the random forest model as an example, sample points can be obtained through field sampling, and then the samples can be evenly divided into two parts, 50% for training the model and 50% for verification. Such a division strategy helps to ensure that the model can show good generalization performance on different data sets, and through monitoring and adjustment of the verification set, the application effect of the model in actual scenarios can be improved.

[0106] It can be understood that based on the object classification model, the object type in each area can be identified according to the simulated hyperspectral data, and then the distribution positions of various objects in the target area can be obtained according to the actual geographic coordinates corresponding to the spectral data, that is, the object distribution of various objects in the target area can be output.

[0107] For example, lotus is distributed in the area coordinate range B and area range C, and calamus is distributed in the area coordinate range B and area range D. The embodiment of the present application does not limit the form of the distribution data.

[0108] In one embodiment, the present application can determine the state of the wetland according to the distribution of land objects. Specifically, the above steps S201-S203 can be performed periodically to obtain land object distribution data of each period.

[0109] Specifically, the feature distribution data of each period can be compared to determine the changing trend of the feature types on the wetland.

[0110] For example, wetland X can be observed every quarter, and the land feature distribution data in summer and autumn can be compared. It is not difficult to find that the water distribution area is reduced, and then the water area is calculated to be smaller. It can be judged that the lotus population is most widely distributed in summer and the coverage is reduced in autumn.

[0111] The embodiment of the present application can also determine whether the wetland has soil erosion, whether algal blooms and other hazards occur, changes in vegetation and water area, etc. based on the distribution data of each period measured, but the embodiment of the present application is not limited to this.

[0112] The following are device embodiments of the present application, which can be used to execute the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

[0113] See next Figure 4 , is a schematic diagram of the structure of a wetland observation device based on band simulation provided by an exemplary embodiment of the present application. The device can be implemented as all or part of a terminal through software, hardware, or a combination of both, and can also be integrated on a server as an independent module. The wetland observation device based on band simulation 40 in the embodiment of the present application includes a spectral data acquisition module 410, a band simulation module 420, and a land object classification module 430.

[0114] The spectral data acquisition module 410 is used to acquire satellite multispectral data of the target area and acquire UAV hyperspectral data of a preset number of sample areas;

[0115] The band simulation module 420 is used to input the satellite multispectral data and the UAV hyperspectral data into the trained band simulation model, so that the trained band simulation model simulates the mapping relationship between the UAV hyperspectral data in each sample area and the satellite multispectral data in the corresponding area to obtain the simulated hyperspectral data of the target area;

[0116] The object classification module 430 is used to input the simulated hyperspectral data into the trained object classification model, so that the trained object classification model analyzes the simulated hyperspectral data based on the characteristic bands and outputs the object distribution data of the target area;

[0117] Wherein, each of the sample areas does not overlap with each other and is located within the range of the target area.

[0118] It should be noted that the device 40 provided in the above embodiment only uses the division of the above functional modules as an example when executing the wetland observation method based on band simulation. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device provided in the above embodiment and the wetland observation method embodiment based on band simulation belong to the same concept. The implementation process is detailed in the method embodiment and will not be repeated here.

[0119] An embodiment of the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above-mentioned method embodiments when executing the program.

[0120] See also Figure 5 , which is a structural block diagram of an electronic device provided in an embodiment of the present application.

[0121] like Figure 5 As shown, the electronic device 500 includes: a processor 501 and a memory 502 .

[0122] In the embodiment of the present application, the processor 501 is the control center of the computer system, which can be a processor of a physical machine or a processor of a virtual machine. The processor 501 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 501 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array).

[0123] The processor 501 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also called a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state.

[0124] The memory 502 may include one or more computer-readable storage media, which may be non-transitory. The memory 502 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments of the present application, the non-transitory computer-readable storage medium in the memory 502 is used to store at least one instruction, which is used to be executed by the processor 501 to implement the method in the embodiment of the present application.

[0125] In some embodiments, the electronic device 500 further includes: a peripheral device interface 503 and at least one peripheral device. The processor 501, the memory 502 and the peripheral device interface 503 can be connected via a bus or a signal line. Each peripheral device can be connected to the peripheral device interface 503 via a bus, a signal line or a circuit board. Specifically, the peripheral devices include: a display screen 504, a camera 505 and an audio circuit 506. The peripheral device interface 503 can be used to connect at least one peripheral device related to I / O (Input / Output) to the processor 501 and the memory 502.

[0126] In some embodiments of the present application, the processor 501, the memory 502, and the peripheral device interface 503 are integrated on the same chip or circuit board; in some other embodiments of the present application, any one or two of the processor 501, the memory 502, and the peripheral device interface 503 can be implemented on a separate chip or circuit board. This embodiment of the present application does not specifically limit this.

[0127] The display screen 504 is used to display a UI. The UI may include graphics, text, icons, videos, and any combination thereof. When the display screen 504 is a touch display screen, the display screen 504 also has the ability to collect touch signals on the surface or above the surface of the display screen 504. The touch signal may be input as a control signal to the processor 501 for processing. At this time, the display screen 504 may also be used to provide virtual buttons and / or virtual keyboards, also known as soft buttons and / or soft keyboards.

[0128] In some embodiments of the present application, the display screen 504 may be one, which is arranged on the front panel of the electronic device 500; in other embodiments of the present application, the display screen 504 may be at least two, which are arranged on different surfaces of the electronic device 500 or are folded; in still other embodiments of the present application, the display screen 504 may be a flexible display screen, which is arranged on a curved surface or a folded surface of the electronic device 500. Even more, the display screen 504 may be arranged in a non-rectangular irregular shape, that is, a special-shaped screen. The display screen 504 may be made of materials such as LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode) and the like.

[0129] The camera 505 is used to capture images or videos. Optionally, the camera 505 includes a front camera and a rear camera. Usually, the front camera is arranged on the front panel of the electronic device, and the rear camera is arranged on the back of the electronic device. In some embodiments, there are at least two rear cameras, which are any one of a main camera, a depth of field camera, a wide-angle camera, and a telephoto camera, so as to realize the fusion of the main camera and the depth of field camera to realize the background blur function, the fusion of the main camera and the wide-angle camera to realize the panoramic shooting and VR (Virtual Reality) shooting function or other fusion shooting functions. In some embodiments of the present application, the camera 505 may also include a flash. The flash can be a monochrome temperature flash or a dual-color temperature flash. The dual-color temperature flash refers to a combination of a warm light flash and a cold light flash, which can be used for light compensation at different color temperatures.

[0130] The audio circuit 506 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, and convert the sound waves into electrical signals and input them into the processor 501 for processing. For the purpose of stereo sound collection or reduction, there may be multiple microphones, which are respectively arranged at different parts of the electronic device 500. The microphone may also be an array microphone or an omnidirectional collection microphone.

[0131] The power supply 507 is used to power various components in the electronic device 500. The power supply 507 can be an alternating current, a direct current, a disposable battery, or a rechargeable battery. When the power supply 507 includes a rechargeable battery, the rechargeable battery can be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery charged through a wired line, and a wireless rechargeable battery is a battery charged through a wireless coil. The rechargeable battery can also be used to support fast charging technology.

[0132] The electronic device structure block diagram shown in the embodiment of the present application does not constitute a limitation on the electronic device 500. The electronic device 500 may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0133] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the method of any of the above embodiments are implemented. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a micro drive, and a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0134] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution can be essentially or in other words, the part that contributes to the relevant technology can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A wetland observation method based on band simulation, characterized in that: include: Obtain satellite multispectral data of the target area and obtain drone hyperspectral data of a preset number of sample areas; The satellite multispectral data and the UAV hyperspectral data are input into a trained band simulation model, so that the trained band simulation model simulates and obtains the simulated hyperspectral data of the target area according to the mapping relationship between the UAV hyperspectral data in each sample area and the satellite multispectral data in the corresponding area; Inputting the simulated hyperspectral data into a trained land object classification model, so that the trained land object classification model analyzes the simulated hyperspectral data based on characteristic bands and outputs land object distribution data of the target area; Wherein, each of the sample areas does not overlap with each other and is located within the range of the target area; After obtaining the satellite multispectral data of the target area and obtaining the drone hyperspectral data of a preset number of sample areas, the method further includes: Preprocessing the satellite multispectral data and the drone hyperspectral data, and aligning the satellite multispectral data with the drone hyperspectral data of the preset number of sample areas based on geographic location information; Performing radiation correction and atmospheric correction on the aligned data, and outputting the corrected data; The step of inputting the satellite multispectral data and the drone hyperspectral data into a trained band simulation model comprises: Inputting the corrected data into a trained band simulation model; The trained band simulation model is obtained by training according to the following steps, including: The drone hyperspectral data of a preset area and the satellite multispectral data of the preset area are used as a sample set; The sample set is input into the band simulation model, each band of the satellite multispectral data in the preset area is used as an independent variable, and each band of the drone hyperspectral data in the preset area is used as a dependent variable, so that the band simulation model learns the mapping relationship between the reflectivity of the drone hyperspectral data in the preset area and the reflectivity of the satellite multispectral data; Outputting the trained band simulation model; Before obtaining the drone hyperspectral data of a preset number of sample areas, the method further includes: estimating the types of objects in the target area and the estimated distribution data of various objects in the target area; Selecting a preset number of sample areas within the target area based on the estimated distribution data; wherein each of the sample areas contains at least one of the ground objects, and the set of ground object types contained in all the sample areas includes all ground object types in the target area; After obtaining the satellite multispectral data of the target area and before obtaining the drone hyperspectral data of a preset number of sample areas, the method further includes: Determine the spectral reflectance distribution in the target area according to the satellite multispectral data; The area with the lowest spectral reflectance and the area with the highest spectral reflectance in the target area are determined according to the spectral reflectance distribution, and the corresponding area is determined as the sample area.

2. The method according to claim 1, characterized in that Before inputting the simulated hyperspectral data into the trained ground object classification model, the method further includes: Obtain the reflectivity of each ground feature in each band; Obtain the reflectivity difference of various objects in the same band; The characteristic band for identifying the type of the land feature is determined based on the reflectivity and the reflectivity difference.

3. The method according to claim 1, characterized in that The step of acquiring satellite multispectral data of the target area and acquiring drone hyperspectral data of a preset number of sample areas is performed at intervals of a preset time period; After outputting the object distribution data of the target area, the method further includes: The change trend of the ground object distribution data of the target area is calculated based on the ground object distribution data of the target area in each of the preset time periods.

4. A wetland observation device based on band simulation, characterized in that: include: The spectral data acquisition module is used to acquire satellite multispectral data of the target area and obtain UAV hyperspectral data of a preset number of sample areas; A band simulation module is used to input the satellite multispectral data and the UAV hyperspectral data into a trained band simulation model, so that the trained band simulation model simulates and obtains the simulated hyperspectral data of the target area according to the mapping relationship between the UAV hyperspectral data in each sample area and the satellite multispectral data in the corresponding area; A ground object classification module, used for inputting the simulated hyperspectral data into a trained ground object classification model, so that the trained ground object classification model analyzes the simulated hyperspectral data based on characteristic bands and outputs ground object distribution data of the target area; Wherein, each of the sample areas does not overlap with each other and is located within the range of the target area; After obtaining the satellite multispectral data of the target area and obtaining the drone hyperspectral data of a preset number of sample areas, the method further includes: Preprocessing the satellite multispectral data and the drone hyperspectral data, and aligning the satellite multispectral data with the drone hyperspectral data of the preset number of sample areas based on geographic location information; Performing radiation correction and atmospheric correction on the aligned data, and outputting the corrected data; The step of inputting the satellite multispectral data and the drone hyperspectral data into a trained band simulation model comprises: Inputting the corrected data into a trained band simulation model; The trained band simulation model is obtained by training according to the following steps, including: The drone hyperspectral data of a preset area and the satellite multispectral data of the preset area are used as a sample set; The sample set is input into the band simulation model, each band of the satellite multispectral data in the preset area is used as an independent variable, and each band of the drone hyperspectral data in the preset area is used as a dependent variable, so that the band simulation model learns the mapping relationship between the reflectivity of the drone hyperspectral data in the preset area and the reflectivity of the satellite multispectral data; Outputting the trained band simulation model; Before obtaining the drone hyperspectral data of a preset number of sample areas, the method further includes: estimating the types of objects in the target area and the estimated distribution data of various objects in the target area; Selecting a preset number of sample areas within the target area based on the estimated distribution data; wherein each of the sample areas contains at least one of the ground objects, and the set of ground object types contained in all the sample areas includes all ground object types in the target area; After obtaining the satellite multispectral data of the target area and before obtaining the drone hyperspectral data of a preset number of sample areas, the method further includes: Determine the spectral reflectance distribution in the target area according to the satellite multispectral data; The area with the lowest spectral reflectance and the area with the highest spectral reflectance in the target area are determined according to the spectral reflectance distribution, and the corresponding area is determined as the sample area.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images