A method for inverting satellite image wetland biomass based on unmanned aerial vehicle data

Abstract

This application discloses a method for retrieving wetland biomass from satellite imagery based on UAV data, relating to the field of remote sensing information processing technology. The method includes: acquiring satellite imagery of the target study area; preprocessing the satellite imagery; calculating the broadband vegetation index distribution of the satellite imagery based on the preprocessed satellite imagery; inputting the broadband vegetation index distribution into a trained wetland biomass retrieval model; retrieving the wetland biomass of each pixel based on the broadband vegetation index in the satellite imagery; and outputting a wetland biomass distribution map of the target study area. This application, by combining high-resolution satellite data, establishes a multivariate wetland biomass retrieval model, improving the accuracy of directly retrieving vegetation biomass from satellite imagery. This can enhance the accuracy of data obtained from wetland vegetation biomass monitoring and provide data support for wetland vegetation parameter retrieval, environmental monitoring, and carbon sequestration research.

Description

Technical Field

[0001] This application relates to the field of remote sensing information processing technology, and in particular to a method for retrieving wetland biomass from satellite imagery based on UAV data. Background Technology

[0002] Wetlands are vital ecosystems formed by the interaction of land and water. However, wetlands are susceptible to change due to climate change and human activities, with some wetlands transforming from carbon sinks to carbon sources. Therefore, wetland protection and restoration are crucial. Aboveground biomass (AGB) is the organic matter produced by plant photosynthetic activity. Wetland vegetation AGB is a key indicator for assessing the health status and carbon storage of wetland ecosystems, and also an important parameter for assessing wetland ecosystem assets. Estimating wetland vegetation biomass can provide a scientific basis for the protection and restoration of wetland ecosystems in order to achieve carbon neutrality goals.

[0003] In wetland biomass measurement methods using relevant technologies, it is generally necessary to establish multiple quadrats in the field, weigh and statistically analyze the quadrats, and then deduce the plant biomass of the entire wetland. Due to the complex and variable nature of wetland environments and poor accessibility, traditional manual sampling methods for obtaining aboveground biomass are not only time-consuming and labor-intensive but also difficult to implement on a large scale. Related technologies can also monitor wetland changes on a large scale in space and time using remote sensing satellites. Advances in remote sensing technology have provided a practical and economical method for monitoring wetland changes and mapping vegetation physiological parameters. Although this method can provide an effective data source for large-scale estimation of aboveground biomass, the accuracy of the calculated biomass is limited due to the low spatial and temporal resolution of satellite remote sensing imagery and cloud interference, and it cannot accurately reflect the biomass of the entire wetland.

[0004] Therefore, there is currently a lack of a method that can quickly, efficiently, and accurately measure wetland biomass. Summary of the Invention

[0005] This application provides a method for retrieving wetland biomass from satellite imagery based on UAV data, to address the shortcomings of the aforementioned related technologies. The technical solution is as follows:

[0006] In a first aspect, embodiments of this application provide a method for retrieving wetland biomass from satellite imagery based on UAV data, characterized in that it includes:

[0007] Acquire satellite imagery of the target study area;

[0008] The satellite imagery is preprocessed, and the broadband vegetation index distribution of the satellite imagery is calculated based on the preprocessed satellite imagery.

[0009] The broadband vegetation index distribution is input into the trained wetland biomass inversion model, and the wetland biomass of the corresponding pixel is obtained by inverting the broadband vegetation index of each pixel in the satellite image through the wetland biomass inversion model.

[0010] Output a wet biomass distribution map of the target study area;

[0011] The trained wetland biomass inversion model is trained based on UAV point cloud data, UAV hyperspectral data and measured wetland biomass in the sample area of ​​the target study area.

[0012] In one alternative embodiment of the first aspect, the trained wetland biomass inversion model is trained based on the following steps:

[0013] Select a preset sub-region within the target study area;

[0014] Acquire UAV point cloud data and UAV hyperspectral data of the preset sub-region;

[0015] Collect measured wet biomass of the sample area within the preset sub-region;

[0016] Using the measured wet biomass as the response variable and the UAV point cloud data and UAV hyperspectral data in the sample area as prediction variables, the wetland biomass inversion model is trained based on the response variable and the prediction variable, so that the wetland biomass inversion model learns the functional relationship between the prediction variable and the response variable.

[0017] The trained wetland biomass inversion model is obtained.

[0018] In one alternative embodiment of the first aspect, after acquiring the UAV point cloud data and UAV hyperspectral data of the preset sub-region, the method further includes:

[0019] The UAV point cloud data and UAV hyperspectral data are processed into data with the same spatial resolution based on the nearest neighbor resampling method, and the radiation value of the collected data is adjusted by radiometric correction so that the adjusted radiation value matches the actual surface reflectance in the preset sub-region.

[0020] The method of using the measured wet biomass as the response variable and the UAV point cloud data and UAV hyperspectral data within the sample area as prediction variables includes:

[0021] The vegetation height distribution within the preset sub-region is calculated based on the corrected UAV point cloud data;

[0022] The narrow-band vegetation index distribution within the preset sub-region is obtained based on the corrected UAV hyperspectral data processing.

[0023] The vegetation height in each pixel is obtained based on the vegetation height distribution, and the narrowband vegetation index in each pixel is obtained based on the narrowband vegetation index distribution. The vegetation height and narrowband vegetation index in the same pixel are selected as the prediction variables, and the measured wet biomass in the same pixel is selected as the response variable.

[0024] In one alternative embodiment of the first aspect, the calculation of the vegetation height distribution within the preset sub-region based on the corrected UAV point cloud data includes:

[0025] Based on the corrected UAV point cloud data, the digital elevation model and digital surface model of the preset sub-region are obtained respectively, and the vegetation height model of the preset sub-region is constructed. Based on the vegetation height model, the vegetation height data of each pixel in the preset sub-region is determined, and the vegetation height distribution in the preset sub-region is obtained.

[0026] The narrow-band vegetation index distribution within the preset sub-region is obtained by processing the corrected UAV hyperspectral data, including:

[0027] Feature extraction is performed based on the corrected UAV hyperspectral data to obtain the feature bands corresponding to each pixel in the preset sub-region. The corresponding reflectance value is calculated based on each feature band, and the narrow band vegetation index is calculated based on the reflectance value to obtain the narrow band vegetation index distribution in the preset sub-region.

[0028] In one alternative embodiment of the first aspect, the preprocessing of the satellite imagery includes:

[0029] The satellite imagery is subjected to atmospheric correction and then processed to a size that matches the target study area.

[0030] The broadband vegetation index distribution of the satellite imagery calculated based on the preprocessed satellite imagery includes:

[0031] Multiple feature bands are extracted from the preprocessed satellite image, and the reflectance value corresponding to each feature band is calculated. Based on the reflectance value, the broadband vegetation index corresponding to each pixel on the preprocessed satellite image is calculated, and the broadband vegetation index distribution is obtained.

[0032] In one alternative embodiment of the first aspect, the step of obtaining the wet biomass of a corresponding pixel based on the broadband vegetation index of each pixel in the satellite image using the wetland biomass inversion model includes:

[0033] Using the broadband vegetation index of each pixel in the target study area as the prediction variable, the wetland biomass inversion model is used to invert the response variable of each pixel according to the functional relationship between the prediction variable and the response variable, and outputs the wet biomass of each pixel.

[0034] In one alternative embodiment of the first aspect, the method further includes:

[0035] The corresponding measured dry biomass is obtained based on the measured wet biomass of the sample area within the preset sub-region.

[0036] Determine the functional relationship between measured wet biomass and measured dry biomass;

[0037] After outputting the wet biomass distribution map of the target study area, the method further includes:

[0038] Based on the functional relationship between the measured wet biomass and the measured dry biomass, a dry biomass distribution map of the target study area is generated on the basis of the wet biomass distribution map.

[0039] Secondly, embodiments of this application also provide a device for retrieving wetland biomass from satellite imagery based on UAV data, comprising:

[0040] The data acquisition module is used to acquire satellite imagery of the target study area;

[0041] The data processing module is used to preprocess the satellite imagery and calculate the broadband vegetation index distribution of the satellite imagery based on the preprocessed satellite imagery.

[0042] The data analysis module is used to input the broadband vegetation index distribution into the trained wetland biomass inversion model, and to obtain the wetland biomass of the corresponding pixel by inverting the broadband vegetation index of each pixel in the satellite image through the wetland biomass inversion model.

[0043] The data analysis module is also used to output a wet biomass distribution map of the target study area;

[0044] The trained wetland biomass inversion model is trained based on UAV point cloud data, UAV hyperspectral data and measured wetland biomass in the sample area of ​​the target study area.

[0045] Thirdly, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method provided by the first aspect or any implementation thereof of the embodiments of this application.

[0046] Fourthly, this 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 embodiments of this application or any implementation thereof.

[0047] The beneficial effects of the technical solutions provided in some embodiments of this application include at least the following:

[0048] This application provides a method for retrieving wetland biomass from satellite imagery based on UAV data. Addressing the limitations of in-situ wetland biomass samples and the problems of low spatial and spectral resolution, mixed pixels, and lack of three-dimensional vegetation information in satellite imagery, this method introduces UAV hyperspectral and lidar data to retrieve local biomass, expanding the wetland biomass sample and supplementing limited ground sampling data. Furthermore, by combining high-resolution satellite data, a multivariate wetland biomass retrieval model is established, improving the accuracy of directly retrieving vegetation biomass from satellite imagery. Mapping wetland biomass under high-resolution satellite imagery enhances the accuracy of data obtained from wetland vegetation biomass monitoring, providing data support for wetland vegetation parameter retrieval, environmental monitoring, and carbon sequestration research. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating a method for retrieving wetland biomass from satellite imagery based on UAV data, provided in an embodiment of this application.

[0051] Figure 2 This is a schematic diagram of wetland biomass distribution, illustrating a method for retrieving wetland biomass from satellite imagery based on UAV data, provided in an embodiment of this application.

[0052] Figure 3 This is a schematic diagram of the structure of a device for retrieving wetland biomass from satellite images based on UAV data, provided in an embodiment of this application.

[0053] Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0055] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or modules is not limited to the steps or modules listed, but may optionally include steps or modules not listed, or may optionally include other steps or modules inherent to such process, method, product, or apparatus.

[0056] It should be noted that the terms "first" and "second" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects. It is understood that "first" and "second" can be interchanged in a specific order or sequence where permitted. It should be understood that the objects distinguished by "first" and "second" can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in an order other than those described or illustrated herein.

[0057] The present application will now be described in detail with reference to specific embodiments.

[0058] Next, combine Figure 1 This application introduces a method for retrieving wetland biomass from satellite imagery based on UAV data, provided by an embodiment of the present application. For details, please refer to... Figure 1 , Figure 1 This illustration shows a flowchart of a method for retrieving wetland biomass from satellite imagery based on UAV data, provided in an embodiment of this application. Figure 1 As shown, the method includes the following steps:

[0059] S101, Acquire satellite imagery of the target study area;

[0060] S102, preprocess the satellite image, and calculate the broadband vegetation index distribution of the satellite image based on the preprocessed satellite image.

[0061] S103, input the broadband vegetation index distribution into the trained wetland biomass inversion model, and obtain the wetland biomass of the corresponding pixel by inverting the broadband vegetation index of each pixel in the satellite image through the wetland biomass inversion model.

[0062] S104, Output the wet biomass distribution map of the target study area;

[0063] Specifically, in S101, the target study area can be selected as the wetland area to be studied, and satellite images taken by satellites within the coordinate range can be selected.

[0064] Specifically, in situations involving complex environments and small-scale research areas, priority can be given to high spatial resolution satellite data from the peak vegetation growth period, such as the Gaofen series and Ziyuan series of domestic satellites.

[0065] For example, the ZY1F satellite can be used for monitoring and feature capture. The ZY1F image data has a total of 9 bands, one panchromatic band and eight multispectral bands. By fusing the multispectral bands and the panchromatic bands, a high-resolution satellite remote sensing image with a spatial resolution of 2.5m×2.5m can be obtained.

[0066] Specifically, following S101, step S102 is performed to preprocess the satellite imagery, including:

[0067] The satellite imagery is subjected to atmospheric correction and then processed to a size that matches the target study area.

[0068] In some embodiments, atmospheric correction can eliminate the influence of atmospheric scattering and refraction on plant spectra. A simplified and robust surface reflectance estimation method (SREM) based on the 6SV RT model equations is used for atmospheric correction of the image, further improving image quality and interpretability. Satellite imagery can be processed through geometric correction to ensure correspondence between the satellite imagery and actual geospatial coordinates. For example, aligning the resolution and corresponding geographic coordinates of the satellite imagery with the geographic coordinates of the actual target study area can maximize the match between the satellite imagery and the target study area size, thereby reducing biomass inversion errors.

[0069] Furthermore, in S102, the broadband vegetation index distribution of the satellite image can be calculated based on the preprocessed satellite image. Multiple feature bands are extracted based on the preprocessed satellite image, and the reflectance value corresponding to each feature band is calculated. The broadband vegetation index corresponding to each pixel on the preprocessed satellite image is calculated based on the reflectance value, thus obtaining the broadband vegetation index distribution.

[0070] Understandably, a wideband refers to a band with a broader spectral range corresponding to an image. The spectral range of a wideband is greater than that of a narrowband. Taking the ZY1F satellite as an example, the NIR (near-infrared) band is 776-895nm with a bandwidth of 120nm; the red band is 635-694nm with a bandwidth of 60nm. Within the bandwidth corresponding to each characteristic band, there is only a unique reflectance value. Based on this reflectance value, the corresponding wideband vegetation index can be calculated. Examples of typical vegetation indices include the Normalized Difference Vegetation Index (NDVI), Difference Vegetation Index (DVI), Equal Value Vegetation Index (EVI), and Soil-Adjusted Vegetation Index (SAVI). The following formula can be applied:

[0071] RNDVI=(R760-R705) / (R760+R705);

[0072] mRESR=(R760-R445) / (R705+R445);

[0073] ReCI = (R860 / R669) – 1.

[0074] In some embodiments, the process of extracting multiple feature bands includes:

[0075] Hyperspectral data was dimensionality reduced using the original hyperspectral curves and their first derivative curves. Bands with zero first derivatives and peak values ​​were selected to identify multiple characteristic bands. Next, based on the dimensionality-reduced bands, various vegetation indices, including broadband, narrow-band, leaf structure, and pigment indices, were calculated as feature variables. Simultaneously, broadband vegetation indices were calculated based on the bands in the satellite imagery. Then, to reduce the uncertainty between collinear variables during feature selection in the inversion model, Pearson correlation coefficients were used for preliminary screening, eliminating variables with correlation coefficients greater than 0.95 and lower correlations with biomass.

[0076] Therefore, the broadband vegetation index distribution of the target study area can be obtained based on the broadband vegetation index corresponding to each pixel in the satellite image.

[0077] Next, step S103 can be performed to input the broadband vegetation index distribution into the pre-trained wetland biomass inversion model, specifically including:

[0078] Using the broadband vegetation index of each pixel in the target study area as the predictor variable, the wetland biomass inversion model is used to invert the predictor variable corresponding to each pixel according to the functional relationship between the predictor variable and the response variable, and outputs the wet biomass corresponding to each pixel.

[0079] Further, by performing step S104, a wet biomass distribution map of the target study area can be generated based on the biomass corresponding to each pixel.

[0080] For example, such as Figure 2 As shown, taking Qilihai Wetland as an example, Figure 2 This is a map showing the distribution of wet biomass in the Qilihai Wetland.

[0081] In some embodiments, the wetland biomass inversion model in S103 is trained based on the following steps:

[0082] S1031, Select a preset sub-region within the target research area;

[0083] S1032, acquire UAV point cloud data and UAV hyperspectral data of the preset sub-region;

[0084] S1033, Collect the measured wet biomass of the sample area within the preset sub-region;

[0085] S1034, using the measured wet biomass as the response variable and the UAV point cloud data and UAV hyperspectral data in the sample area as prediction variables, the wetland biomass inversion model is trained based on the prediction variables and the response variables, so that the wetland biomass inversion model learns the functional relationship between the prediction variables and the response variables.

[0086] S1035, the trained wetland biomass inversion model is obtained.

[0087] Specifically, in S1031, a typical area can be selected as a preset sub-region within the target study area. A typical area can be understood as an area that conforms to the characteristics of biological types within the target study area. For example, some areas may have more herbaceous plants, while some areas may be dominated by fungi. Alternatively, multiple smaller sub-regions can be randomly selected from the wetland coverage area of ​​the target study area. For example, a sub-region within the range of a×b can be selected. The selection of sub-regions is not limited in this embodiment.

[0088] Specifically, in S1032, LiDAR and hyperspectral cameras mounted on DJI drones can be used to acquire information about the target study area. Drones are flexible, efficient, and offer extremely high resolution, overcoming sampling constraints in inaccessible or complex wetland areas and reducing the cost of wetland monitoring. Airborne LiDAR can identify three-dimensional features of vegetation within the wetland, providing vegetation height information. Airborne hyperspectral imaging offers ultra-high spectral resolution, providing detailed spectral information, particularly red-edge information reflecting vegetation growth. Fusing LiDAR and hyperspectral data to obtain rich vegetation information allows for more accurate estimation of wetland vegetation biomass.

[0089] For example, the Zenmuse L1 LiDAR on a DJI drone captures point cloud data at a flight altitude of 100m, a ranging accuracy of 3cm, and an average point density of 50 points / m. 2 A lightweight hyperspectral imager, the ZK-VNIR-FPG480, mounted on a drone, can be used to acquire local hyperspectral images of the study area. The resulting spectral range covers 300 bands from 400nm to 1000nm, with a spectral resolution of 2.8nm. The drone flies at an altitude of 100m and a speed of 7.5m / s, achieving a spatial resolution of 0.15m. Drone imagery can obtain the height of wetland vegetation and more detailed spectral information, contributing to improved accuracy in wetland biomass estimation.

[0090] In some embodiments, after acquiring the UAV point cloud data and UAV hyperspectral data of the preset sub-region in S1032, preprocessing can be performed on the UAV point cloud data and UAV hyperspectral data, specifically including:

[0091] The UAV point cloud data and UAV hyperspectral data are processed into data with the same spatial resolution using the nearest neighbor resampling method, and the radiation value of the collected data is adjusted by radiometric correction so that the adjusted radiation value matches the actual surface reflectance in the preset sub-region.

[0092] For example, geometric correction can ensure the geospatial correspondence between satellite and UAV imagery. Furthermore, UAV hyperspectral imagery and LiDAR data can be resampled to the same spatial resolution using the nearest neighbor method, ensuring the continuity and spatial consistency of multi-source data. Secondly, radiometric correction adjusts the radiance values ​​acquired by the sensors to match the actual surface reflectance. Image preprocessing ensures the quality and consistency of multi-source remote sensing data from satellites, UAVs, and other sources, and lays a data foundation for subsequent vegetation index calculations and biomass inversion, ensuring the accuracy and feasibility of the research.

[0093] In S1034, the step of using the measured wet biomass as the response prediction variable and the UAV point cloud data and UAV hyperspectral data within the sample area as the prediction response variables includes:

[0094] The vegetation height distribution within the preset sub-region is calculated based on the corrected UAV point cloud data;

[0095] The narrow-band vegetation index distribution within the preset sub-region is obtained based on the corrected UAV hyperspectral data processing.

[0096] The vegetation height in each pixel is obtained based on the vegetation height distribution, and the narrowband vegetation index in each pixel is obtained based on the narrowband vegetation index distribution. The vegetation height and narrowband vegetation index in the same pixel are selected as the response prediction variables, and the measured wet biomass in the same pixel is selected as the predicted response variable.

[0097] Specifically, based on UAV point cloud data, i.e. LiDAR data, after ground point classification, a digital surface model (DSM) and a digital elevation model (DEM) can be calculated and obtained, thereby establishing a vegetation height model (CHM) and obtaining vegetation height data within the study area, i.e., the vegetation height distribution within a preset sub-region.

[0098] It should be noted that, for narrow-band indices, taking the hyperspectral data involved in this application as an example, the 760-906nm ​​band is near-infrared, containing multiple reflectance values ​​corresponding to wavelengths such as 760, 775, 832, 860, and 906nm. The 699-733nm band is the red-edge band, including reflectance values ​​corresponding to wavelengths such as 699, 705, 720, and 733nm. The red band includes values ​​at 650, 669, and 677nm. The reflectance values ​​of the corresponding NIR (near-infrared), red-edge, and red bands can be selected according to different vegetation index formulas and substituted into the calculation formulas of various typical vegetation indices such as Normalized Difference Vegetation Index (NDVI), Difference Vegetation Index (DVI), Equal Value Vegetation Index (EVI), and Soil Adjusted Vegetation Index (SAVI) to calculate the narrow-band vegetation index. The wavelength range corresponding to the narrow band is smaller than that of the wide band.

[0099] In some embodiments, during model training, a random forest biomass estimation model based on UAV imagery can be constructed. Measured biomass data is used as the response variable, while plant height, characteristic bands of the UAV imagery, and vegetation indices are used as predictor variables, thus initially constructing a random forest inversion model for typical wetland vegetation biomass. The feature parameters are optimized using a particle swarm optimization algorithm. During model execution, the ratio of the training set to the validation set is set to 7:3. Since the randomness of the random forest will produce different results in each execution of the algorithm, a preset number of runs are performed for each model, and the average value is taken to ensure the robustness of the model. Finally, the estimation accuracy of the model is evaluated using R² and RMSE. The average R² of the biomass model based on UAV inversion is 0.85, and the average RMSE is 0.65.

[0100] In some embodiments, in S1033, a 1m×1m sample plot can be set up and its geographical coordinates recorded. Then, all the above-ground plants in the sample plot are harvested and weighed using an electronic scale with an accuracy of 5g to obtain the vegetation wet biomass.

[0101] In some embodiments, a portion of the vegetation wet biomass obtained by weighing can be randomly selected and weighed, and the dry weight of the sample can be measured after drying it to constant weight at 75°C for 48 hours in the laboratory. The functional relationship between the measured wet biomass and the measured dry biomass can be determined.

[0102] Therefore, after outputting the wet biomass distribution map of the target study area in S104, the dry biomass distribution map of the target study area can be generated based on the functional relationship between the measured wet biomass and the measured dry biomass.

[0103] The following are apparatus embodiments of this application, which can be used to execute the method embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the method embodiments of this application.

[0104] Please see below. Figure 3 This is a schematic diagram of a device for retrieving wetland biomass from satellite imagery based on UAV data, provided as an exemplary embodiment of this application. This device can be implemented as all or part of a terminal through software, hardware, or a combination of both, or it can be integrated as an independent module on a server. The device for retrieving wetland biomass from satellite imagery based on UAV data in this embodiment can be applied to a terminal or the cloud. The device 30 includes a data acquisition module 301, a data processing module 302, and a data analysis module 303, wherein:

[0105] Data acquisition module 301 is used to acquire satellite imagery of the target study area;

[0106] The data processing module 302 is used to preprocess the satellite image and calculate the broadband vegetation index distribution of the satellite image based on the preprocessed satellite image.

[0107] The data analysis module 303 is used to input the broadband vegetation index distribution into the trained wetland biomass inversion model, and to obtain the wetland biomass of the corresponding pixel by inverting the broadband vegetation index of each pixel in the satellite image through the wetland biomass inversion model.

[0108] The data analysis module 303 is also used to output a wet biomass distribution map of the target study area;

[0109] The trained wetland biomass inversion model is trained based on UAV point cloud data, UAV hyperspectral data and measured wetland biomass in the sample area of ​​the target study area.

[0110] It should be noted that the apparatus 30 provided in the above embodiments, when executing a method for retrieving wetland biomass from satellite imagery based on UAV data, is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the apparatus provided in the above embodiments and the method embodiment for retrieving wetland biomass from satellite imagery based on UAV data belong to the same concept, and their implementation process is detailed in the method embodiment, which will not be repeated here.

[0111] This 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 executes the program to implement the steps of any of the methods described above.

[0112] Please see Figure 4 This is a structural block diagram of an electronic device provided in an embodiment of this application.

[0113] like Figure 4 As shown, the electronic device 400 includes a processor 401 and a memory 402.

[0114] In this embodiment, the processor 401 is the control center of the computer system, and can be a processor of a physical machine or a processor of a virtual machine. The processor 401 may include one or more processing cores, such as a 4-core processor or an 8-core processor. The processor 401 can be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array).

[0115] Processor 401 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake-up state, also known as a CPU (Central Processing Unit). The coprocessor is a low-power processor used to process data in the standby state.

[0116] Memory 402 may include one or more computer-readable storage media, which may be non-transitory. Memory 402 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments of this application, the non-transitory computer-readable storage media in memory 402 is used to store at least one instruction, which is executed by processor 401 to implement the method in the embodiments of this application.

[0117] In some embodiments, the electronic device 400 further includes a peripheral device interface 403 and at least one peripheral device 404. The processor 401, memory 402, and peripheral device interface 403 can be connected via a bus or signal line. Each peripheral device 404 can be connected to the peripheral device interface 403 via a bus, signal line, or circuit board. Specifically, the peripheral device 404 includes: a display screen, a camera, and audio circuitry. The peripheral device interface 403 can be used to connect at least one I / O (Input / Output) related peripheral device to the processor 401 and memory 402.

[0118] In some embodiments of this application, the processor 401, memory 402, and peripheral device interface 403 are integrated on the same chip or circuit board; in other embodiments of this application, any one or two of the processor 401, memory 402, and peripheral device interface 403 can be implemented on separate chips or circuit boards. This application does not specifically limit the implementation in this regard.

[0119] The block diagram of the electronic device shown in the embodiments of this application does not constitute a limitation on the electronic device 400. The electronic device 400 may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0120] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the methods in any of the foregoing embodiments. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0121] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of software products. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for retrieving wetland biomass from satellite imagery based on UAV data, characterized in that, include: Acquire satellite imagery of the target study area; The satellite imagery is preprocessed, and the broadband vegetation index distribution of the satellite imagery is calculated based on the preprocessed satellite imagery. The broadband vegetation index distribution is input into the trained wetland biomass inversion model, and the wetland biomass of the corresponding pixel is obtained by inverting the broadband vegetation index of each pixel in the satellite image through the wetland biomass inversion model. Output a wet biomass distribution map of the target study area; The trained wetland biomass inversion model is trained based on UAV point cloud data, UAV hyperspectral data, and measured wetland biomass within a preset sub-region of the target study area, including: Select a preset sub-region within the target study area; Acquire UAV point cloud data and UAV hyperspectral data of the preset sub-region; Collect measured wet biomass of the sample area within the preset sub-region; Vegetation height distribution was calculated based on UAV point cloud data, and narrowband vegetation index distribution was obtained based on UAV hyperspectral data processing. Using the measured wet biomass as the response variable and the UAV point cloud data and UAV hyperspectral data in the preset sub-region as the prediction variables, the wetland biomass inversion model is trained based on the prediction variables and the response variables, so that the wetland biomass inversion model learns the functional relationship between the prediction variables and the response variables. The trained wetland biomass inversion model is obtained; The preprocessing of the satellite imagery includes: The satellite imagery is subjected to atmospheric correction and then processed to a size that matches the target study area. The broadband vegetation index distribution of the satellite imagery calculated based on the preprocessed satellite imagery includes: Based on the preprocessed satellite image, the hyperspectral data is dimensionality reduced by using the original hyperspectral curve and the corresponding first derivative curve. Bands with zero first derivative and peak values ​​are selected, and multiple feature bands are extracted. The reflectance value corresponding to each feature band is calculated. Based on the reflectance value, the broadband vegetation index corresponding to each pixel on the preprocessed satellite image is calculated, and the broadband vegetation index distribution is obtained.

2. The method for retrieving wetland biomass from satellite imagery based on UAV data according to claim 1, characterized in that, After acquiring the UAV point cloud data and UAV hyperspectral data of the preset sub-region, the process further includes: The UAV point cloud data and UAV hyperspectral data are processed into data with the same spatial resolution based on the nearest neighbor resampling method, and the radiometric values ​​of the collected data are adjusted by radiometric correction so that the adjusted radiometric values ​​are consistent with the actual surface reflectance in the preset sub-region. The method of using the measured wet biomass as the response variable and the UAV point cloud data and UAV hyperspectral data within the sample area as prediction variables includes: The vegetation height distribution within the preset sub-region is calculated based on the corrected UAV point cloud data; The narrow-band vegetation index distribution within the preset sub-region is obtained based on the corrected UAV hyperspectral data processing. The vegetation height in each pixel is obtained based on the vegetation height distribution, and the narrowband vegetation index in each pixel is obtained based on the narrowband vegetation index distribution. The vegetation height and narrowband vegetation index in the same pixel are selected as the prediction variables, and the measured wet biomass in the same pixel is selected as the response variable.

3. The method for retrieving wetland biomass from satellite imagery based on UAV data according to claim 2, characterized in that, The calculation of vegetation height distribution within the preset sub-region based on corrected UAV point cloud data includes: Based on the corrected UAV point cloud data, the digital elevation model and digital surface model of the preset sub-region are obtained respectively, and the vegetation height model of the preset sub-region is constructed. Based on the vegetation height model, the vegetation height data of each pixel in the preset sub-region is determined, and the vegetation height distribution in the preset sub-region is obtained. The narrow-band vegetation index distribution within the preset sub-region is obtained by processing the corrected UAV hyperspectral data, including: Feature extraction is performed based on the corrected UAV hyperspectral data to obtain the feature bands corresponding to each pixel in the preset sub-region. The corresponding reflectance value is calculated based on each feature band, and the narrow band vegetation index is calculated based on the reflectance value to obtain the narrow band vegetation index distribution in the preset sub-region.

4. The method for retrieving wetland biomass from satellite imagery based on UAV data according to claim 1, characterized in that, The process of obtaining the wet biomass of a corresponding pixel through the wetland biomass inversion model based on the broadband vegetation index of each pixel in the satellite imagery includes: Using the broadband vegetation index of each pixel in the target study area as the prediction variable, the wetland biomass inversion model is used to invert the response variable of each pixel pixel by pixel according to the functional relationship between the prediction variable and the response variable, and outputs the wetland biomass of each pixel.

5. The method for retrieving wetland biomass from satellite imagery based on UAV data according to claim 1, characterized in that, The method further includes: The corresponding measured dry biomass is obtained based on the measured wet biomass of the sample area within the preset sub-region. Determine the functional relationship between measured wet biomass and measured dry biomass; After outputting the wet biomass distribution map of the target study area, the method further includes: Based on the functional relationship between the measured wet biomass and the measured dry biomass, a dry biomass distribution map of the target study area is generated on the basis of the wet biomass distribution map.

6. An apparatus for retrieving wetland biomass from satellite imagery based on UAV data according to any one of claims 1-5, characterized in that, include: The data acquisition module is used to acquire satellite imagery of the target study area; The data processing module is used to preprocess the satellite imagery and calculate the broadband vegetation index distribution of the satellite imagery based on the preprocessed satellite imagery. The data analysis module is used to input the broadband vegetation index distribution into the trained wetland biomass inversion model, and to obtain the wetland biomass of the corresponding pixel by inverting the broadband vegetation index of each pixel in the satellite image through the wetland biomass inversion model. The data analysis module is also used to output a wet biomass distribution map of the target study area; The trained wetland biomass inversion model is trained based on UAV point cloud data, UAV hyperspectral data and measured wetland biomass in the sample area of ​​the target study area.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method as described in any one of claims 1 to 5.

8. 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, it implements the steps of the method as described in any one of claims 1 to 5.

Images