Methods, apparatus, computing devices, and storage media for determining grazing intensity

Abstract

The application discloses a method and device for determining grazing intensity, a computing device and a storage medium. The method for determining grazing intensity is executed in the computing device, and comprises the following steps: acquiring at least one set of historical remote sensing data of permanent grassland in a measured area in a thermal growing season; determining feature data of the remote sensing data by using the acquired sets of historical remote sensing data; inputting the feature data into a grazing probability prediction model for processing to predict the grazing probability of the measured area; and determining the grazing intensity of the measured area based on the grazing probability. The method acquires the grazing probability of the measured area through historical remote sensing data of the measured area, and then obtains the grazing intensity through the grazing probability. The method realizes high-resolution, long-time-series quantitative grazing intensity monitoring of a large-area measured area, and is flexible in extrapolation and high in precision.

Description

Technical Field

[0001] This invention relates to the field of computing device technology, and specifically to a method, apparatus, computing device, and storage medium for determining grazing intensity. Background Technology

[0002] Grazing intensity (GI), also known as grazing pressure, refers to the number of livestock grazing per unit area of ​​grassland within a certain period of time. GI reflects the complex and profound impact of livestock on grassland during grazing, and it is a comprehensive indicator for assessing grassland quality and management level.

[0003] Currently, traditional methods for determining grassland geographic information (GI) include field measurements or statistics, with statistics typically based on spatialization of questionnaire survey information or government census data. However, these methods cannot provide large-area, pixel-level spatiotemporal information on grazing. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a method, apparatus, computing device, and storage medium for determining grazing intensity that overcomes or at least partially solves the above problems.

[0005] According to one aspect of the present invention, a method for determining grazing intensity is provided, executed in a computing device, the method comprising: acquiring at least one set of historical remote sensing data of permanent grassland in a measured area during the hot growing season; using the acquired sets of historical remote sensing data to determine characteristic data of the remote sensing data; inputting the characteristic data into a grazing probability prediction model for processing to predict the grazing probability of the measured area; and determining the grazing intensity of the measured area based on the grazing probability.

[0006] Optionally, in the method for determining grazing intensity according to the present invention, determining the grazing intensity of the tested area based on the grazing probability includes: if the grazing probability is not less than a first threshold, then determining the grazing intensity of the tested area as heavy grazing; if the grazing probability is less than the first threshold and greater than a second threshold, then determining the grazing intensity of the tested area as moderate grazing; if the grazing probability is not greater than the second threshold, then determining the grazing intensity of the tested area as light grazing.

[0007] Optionally, in the method for determining grazing intensity according to the present invention, the remote sensing data includes at least spectral band data and vegetation index data; and at least one set of historical remote sensing data of permanent grassland in the measured area during the hot growing season is acquired, including: determining the permanent grassland area included in the measured area based on the first remote sensing data of the measured area; determining the hot growing season cycle of the measured area based on the second remote sensing data of the measured area; acquiring Landsat and Sentinel remote sensing data corresponding to the permanent grassland area within the same hot growing season cycle, and using the acquired Landsat and Sentinel remote sensing data to determine the spectral band data and vegetation index data.

[0008] Optionally, in the method for determining grazing intensity according to the present invention, the spectral band data includes at least one of the following indicators: surface reflectance in the red band, green band, blue band, near-infrared band, and short-wavelength infrared band; and the vegetation index data includes at least one of the following indicators: normalized difference vegetation index, enhanced vegetation index, and surface water index.

[0009] Optionally, in the method for determining grazing intensity according to the present invention, acquiring Landsat and Sentinel remote sensing data corresponding to the permanent grassland area, and using the acquired Landsat and Sentinel remote sensing data to determine the spectral band data and vegetation index data, includes: using the Cfmask method to screen out data in the Landsat and Sentinel remote sensing data that do not meet preset conditions; fusing the screened Landsat and Sentinel remote sensing data belonging to the same preset time interval to obtain fused data; for each fused data, extracting the spectral band data of the fused data, and calculating the vegetation index data using the extracted spectral band data.

[0010] Optionally, in the method for determining grazing intensity according to the present invention, the vegetation index data of each fused data is calculated by the following formula:

[0011] in, Normalized Difference Vegetation Index (NDVI) The surface reflectance is in the near-infrared band. This represents the surface reflectance value in the red light band. To enhance the vegetation index, The surface reflectance is in the blue light band. This is a surface water index. This refers to the surface reflectance in the short-wavelength infrared band.

[0012] Optionally, in the method for determining grazing intensity according to the present invention, the characteristic data of the remote sensing data are determined by using the acquired historical remote sensing data, including: dividing the historical remote sensing data of the same year into a remote sensing dataset; and for each remote sensing dataset, calculating the average value of each indicator in the remote sensing dataset as the characteristic data of the corresponding year of the remote sensing dataset.

[0013] Optionally, in the method for determining grazing intensity according to the present invention, the first remote sensing data is MODIS-MCD12Q1 remote sensing data; and based on the first remote sensing data of the measured area, determining the permanent grassland area contained in the measured area includes: extracting a grassland map within a preset time period from the MODIS-MCD12Q1 remote sensing data; and taking the area marked as grassland in the grassland map within the preset time period as the permanent grassland area.

[0014] Optionally, in the method for determining grazing intensity according to the present invention, the second remote sensing data is MODIS-MOD11A2 remote sensing data; and based on the second remote sensing data of the measured area, determining the thermogrowing season cycle of the measured area includes: extracting nighttime temperature time series data from the MODIS-MOD11A2 remote sensing data; determining the first date when the first nighttime minimum temperature is less than 5°C and the second date when the last nighttime maximum temperature is greater than 5°C from the nighttime temperature time series data, and taking the time period between the first date and the second date as the thermogrowing season cycle.

[0015] Optionally, in the method for determining grazing intensity according to the present invention, the method further includes a step of generating a grazing probability prediction model: collecting first sample data indicating a heavily grazing area, second sample data indicating an ungrazing area, and historical remote sensing data of the target area within the target area; constructing training samples using the historical remote sensing data of the target area, and determining label data using the first and second sample data; and training the initial grazing probability prediction model based on the training samples and label data to obtain the grazing probability prediction model.

[0016] According to another aspect of the present invention, an apparatus for determining grazing intensity is provided, residing in a computing device, the apparatus comprising: an acquisition module adapted to acquire at least one set of historical remote sensing data of permanent grassland in a measured area during the hot growing season; a first determination module adapted to determine feature data of the acquired historical remote sensing data using the acquired sets of historical remote sensing data; an input module adapted to input the feature data into a grazing probability prediction model for processing to predict the grazing probability of the measured area; and a second determination module adapted to determine the grazing intensity of the measured area based on the grazing probability.

[0017] According to another aspect of the present invention, a computing device is provided, comprising: at least one processor; and a memory storing program instructions, wherein the program instructions are configured to be executed by the at least one processor, the program instructions including instructions for performing the methods described above.

[0018] According to another aspect of the present invention, a readable storage medium storing program instructions is provided, which, when read and executed by a computing device, causes the computing device to perform the method described above.

[0019] According to the present invention, the grazing probability of the measured area is obtained through historical remote sensing data of the measured area, and the grazing intensity is then obtained through the grazing probability. This achieves high-resolution, long-term quantitative monitoring of grazing intensity in a large measured area, with flexible extrapolation and high accuracy.

[0020] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0021] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic diagram of a computing device 100 according to an embodiment of the present invention is shown; Figure 2 A flowchart of a method 200 for determining grazing intensity according to an embodiment of the present invention is shown; Figure 3 A schematic diagram illustrating the grazing probability analysis of an ungrazed sample (a) and a heavily grazed sample (b) according to an embodiment of the present invention is shown. Figure 4 A schematic diagram of a device 400 for determining grazing intensity according to an embodiment of the present invention is shown. Detailed Implementation

[0022] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0023] With the development of remote sensing technology, remote sensing imagery has become a powerful tool for large-scale, long-term dynamic monitoring of grasslands and is widely used in GI monitoring.

[0024] Remote sensing-based grazing activity monitoring algorithms can be summarized as statistical data spatialization methods, differential simulation methods, and machine learning methods.

[0025] Among them, the spatialization method of statistical data mainly relies on statistical data at the administrative level, but statistical data at the administrative level is difficult to generate pixel-level accurate maps.

[0026] Differential simulation methods are widely used in grassland use intensity studies, such as grazing and mowing detection. These methods use various remote sensing measures to simulate the difference between the potential and actual state of vegetation, which serves as an indicator of grassland use status. However, accurately simulating the potential state of vegetation using these methods remains a challenge.

[0027] The class membership probabilities of machine learning models reveal information about the gradual process of mapping, stochastic behavior, and nonlinear relationships. This approach is effective for mapping grazing pressure in the Central Asian steppes based on random forest (RF) models and ground reference data; however, drawing accurate results from only a few years' worth of samples (e.g., a single year's sample) remains a challenge for mapping the interannual dynamics of GI.

[0028] To address the problems existing in the prior art, the present invention is proposed. One embodiment of the present invention provides a method for determining grazing intensity, which can be executed in a computing device. Figure 1 A structural diagram of a computing device 100 according to an embodiment of the present invention is shown. Figure 1 As shown, in the basic configuration 102, the computing device 100 typically includes system memory 106 and one or more processors 104. Memory bus 108 can be used for communication between processor 104 and system memory 106.

[0029] Depending on the desired configuration, processor 104 can be any type of processor, including but not limited to: microprocessor (µP), microcontroller (µC), digital information processor (DSP), or any combination thereof. Processor 104 may include one or more levels of cache such as L1 cache 110 and L2 cache 112, processor core 114, and registers 116. Example processor core 114 may include an arithmetic logic unit (ALU), floating-point unit (FPU), digital signal processing core (DSP core), or any combination thereof. Example memory controller 118 may be used with processor 104, or in some implementations, memory controller 118 may be an internal part of processor 104.

[0030] Depending on the desired configuration, system memory 106 can be any type of memory, including but not limited to: volatile memory (such as RAM), non-volatile memory (such as ROM, flash memory, etc.), or any combination thereof. Physical memory in a computing device typically refers to volatile RAM, and data on a disk needs to be loaded into physical memory before it can be read by processor 104. System memory 106 may include operating system 120, one or more applications 122, and program data 124. Application 122 is actually a series of program instructions that instruct processor 104 to perform corresponding operations. In some embodiments, application 122 may be arranged to execute instructions on the operating system by one or more processors 104 using program data 124. Operating system 120 may be, for example, Linux, Windows, etc., and includes program instructions for handling basic system services and performing hardware-dependent tasks. Application 122 includes program instructions for implementing various user-desired functions, and application 122 may be, for example, a browser, instant messaging software, software development tools (such as integrated development environments (IDEs), compilers, etc.), but is not limited to these. When application 122 is installed on computing device 100, a driver module can be added to operating system 120.

[0031] When computing device 100 starts up, processor 104 reads and executes program instructions from memory 106 of operating system 120. Application 122 runs on operating system 120, utilizing interfaces provided by operating system 120 and underlying hardware to implement various user-expected functions. When user starts application 122, application 122 is loaded into memory 106, and processor 104 reads and executes program instructions from memory 106 of application 122.

[0032] The computing device 100 also includes a storage device 132, which includes a removable storage device 136 and a non-removable storage device 138, both of which are connected to a storage interface bus 134.

[0033] The computing device 100 may also include an interface bus 140 that facilitates communication from various interface devices (e.g., output devices 142, peripheral interfaces 144, and communication devices 146) to the basic configuration 102 via a bus / interface controller 130. Example output devices 142 include a graphics processing unit 148 and an audio processing unit 150. They may be configured to facilitate communication with various external devices such as displays or speakers via one or more A / V ports 152. Example peripheral interfaces 144 may include a serial interface controller 154 and a parallel interface controller 156, which may be configured to facilitate communication with external devices such as input devices (e.g., keyboards, mice, pens, voice input devices, touch input devices) or other peripherals (e.g., printers, scanners, etc.) via one or more I / O ports 158. Example communication devices 146 may include a network controller 160, which may be arranged to facilitate communication with one or more other computing devices 162 via a network communication link through one or more communication ports 164.

[0034] A network communication link can be an example of a communication medium. A communication medium can typically be embodied in a modulated data signal, such as a carrier wave or other transmission mechanism, and can include any information delivery medium. A “modulated data signal” can be a signal in which one or more of its data sets, or where changes to them can be encoded into the signal, are performed. As a non-limiting example, a communication medium can include wired media such as wired networks or leased lines, and various wireless media such as sound, radio frequency (RF), microwave, infrared (IR), or other wireless media. The term “computer-readable medium” as used herein can include both storage media and communication media.

[0035] The computing device 100 also includes a storage interface bus 134 connected to a bus / interface controller 130. The storage interface bus 134 is connected to a storage device 132 adapted to perform data storage. An example storage device 132 may include a removable storage device 136 (e.g., CD, DVD, USB flash drive, removable hard drive, etc.) and a non-removable storage device 138 (e.g., hard disk drive HDD, etc.).

[0036] In the computing device 100 according to the present invention, application 122 includes a plurality of program instructions for executing method 200.

[0037] Figure 2 A flowchart of a method 200 for determining grazing intensity according to an embodiment of the present invention is shown. Method 200 is adapted to be executed in a computing device (e.g., the aforementioned computing device 100).

[0038] like Figure 2 As shown, the purpose of method 200 is to obtain the grazing probability of the measured area through historical remote sensing data of the measured area, and then obtain the grazing intensity through the grazing probability. This achieves high-resolution, long-time-series quantitative monitoring of grazing intensity in a large-area measured area, and also provides a flexible extrapolation and high-precision method for determining the grazing probability.

[0039] Method 200 begins with step 202, in which at least one set of historical remote sensing data of permanent grassland within the measured area during the hot growing season is acquired. The remote sensing data includes at least spectral band data and vegetation index data of the measured area.

[0040] In this embodiment, historical remote sensing data from the same year are grouped together as a single set. For example, to acquire historical remote sensing data for the measured area from 2015 to 2021, the historical remote sensing data from 2015 are grouped together, the historical remote sensing data from 2016 are grouped together, and so on. In some embodiments, the remote sensing data includes at least spectral band data (SBs) and vegetation index data (VIs). The spectral band data includes at least one of the following indicators: surface reflectance (ρ) in the red band. Red ), Surface reflectance in the green band (ρ) Green ), Surface reflectance in the blue light band (ρ) Blue ), Surface reflectance in the near-infrared band (ρ) NIR ) and surface reflectance (ρ) in the short-wavelength infrared band SWIR Vegetation index data include at least one of the following indicators: Normalized Difference Vegetation Index (DOVI), Enhanced Vegetation Index (EVI), and Surface Water Index (LSWI).

[0041] It is easy to understand that grasslands with different grazing intensities exhibit significant spectral differences. In this application, a remote sensing-based method is used to measure the GI (Geometric Spectrum) to provide key information about vegetation under different grazing conditions.

[0042] Specifically, firstly, based on the first remote sensing data of the measured area, the permanent grassland areas contained within the measured area are determined. In this embodiment, the first remote sensing data is MODIS-MCD12Q1 satellite remote sensing data. MCD12Q1 is a MODIS land cover product generated since 2001 based on Terra and Aqua data, providing annual global land cover information and maps with a spatial resolution of 500 meters. It includes five land cover classification systems; in this embodiment, LC_Type1 (using the International Geosphere-Biosphere Programme-IGBP classification) is selected, and the permanent grassland areas of the study area are extracted using the MCD12Q1 land cover map.

[0043] The process of determining the permanent grassland area included in the surveyed area involves the following two steps: The first step is to extract grassland maps for a preset time period from the MODIS-MCD12Q1 remote sensing data. For example, annual grassland maps from 2010 to 2021 can be extracted from the MCD12Q1 LC_Type1 layer for study purposes, with 2010 to 2021 serving as the preset time period for grassland map extraction, and grassland maps for each year obtained from the MODIS-MCD12Q1 remote sensing data.

[0044] The second step involves designating areas on the grassland map that are consistently marked as grassland within a preset time period as permanent grassland areas. In other words, areas on the grassland map that are continuously marked as grassland within the preset time period are considered permanent grassland areas. In some embodiments, grassland areas on the grassland map can be marked using the LC_Type1 land cover IGBP classification scheme.

[0045] The IGBP classification is a geographical standard used to classify land cover into 17 distinct categories. It is a global classification system for vegetation cover published by the International Ecological Union (IGBP). These 17 different IGBP classifications are shown below: Forests / tree vegetation, shrub vegetation, grasslands / open fields, arid and semi-arid regions, wetlands / swamps, arable farmland, snowfields / glaciers, coastal zones / bathing beaches, sandy beaches, artificial alpine areas, artificial low-altitude areas, artificial intermediate-altitude areas, artificial subalpine areas, artificial low-altitude areas, areas with frequent fog and rain, areas with little fog and rain, areas without fog and rain.

[0046] Using the IGBP classification scheme, grassland areas are marked on each grassland map. By comparing the marked grassland maps, if an area is a grassland area within a preset time period (such as from 2010 to 2021), then this area is designated as a permanent grassland area.

[0047] Then, based on the second remote sensing data of the measured area, the hot growing season cycle of the measured area is determined. In this embodiment, the second remote sensing data is MODIS-MOD11A2 remote sensing data.

[0048] Because temperature limits vegetation productivity, the beginning and end of the growing season can be defined as the first day when the minimum temperature is above and below 5°C, respectively. Specifically, nighttime temperature time series data are extracted from MODIS-MOD11A2 remote sensing data. A time series (or dynamic series) refers to a sequence of values ​​of the same statistical indicator arranged in chronological order of their occurrence. In this embodiment, the nighttime temperature time series data refers to the highest and lowest nighttime temperatures for each day of each year (e.g., 2015 to 2021).

[0049] By identifying the first date when the first nighttime minimum temperature falls below 5°C and the second date when the last nighttime maximum temperature exceeds 5°C from nighttime temperature time-series data, the time period between these two dates is defined as the thermotropical growth season cycle. This pattern can be achieved at a pixel scale with a spatial resolution of 1 km using the MODI11A2 nighttime time-series dataset.

[0050] For example, the time span between the first day each year (2015-2021) when the nighttime LST (thermal growth season) is above or below 5°C is used as the thermal growth season cycle for the Hulunbuir Grassland. Specifically, the nighttime thermal growth season can be calculated using land surface temperature and emissivity data of the Hulunbuir Grassland from 2015 to 2021 (i.e., MOD11A2 remote sensing data). The numerical values ​​(DN) of MOD11A2 are converted to values ​​with percentage units according to the following equation.

[0051]

[0052] Preferably, before performing surface temperature calculations, the MOD11A2 remote sensing data can be processed by correcting for noise from cloud contamination, zenith angle variations, and topographic differences, and time series gaps can be filled using linear interpolation.

[0053] Finally, for permanent grassland areas within the same hot growing season cycle, Landsat and Sentinel remote sensing data were acquired for those areas. The acquired Landsat and Sentinel data were then used to determine spectral band data and vegetation index data. Landsat satellite data included all Tire-1 SR images collected by the Google Earth Engine platform, which has a spatial resolution of 30 meters and a temporal resolution of 16 days. Sentinel-2 SR data provided 5-day interval observations at a spatial resolution of 10 meters.

[0054] Specifically, the Cfmask method is used to filter out data from Landsat and Sentinel remote sensing data that do not meet preset conditions. These preset conditions include requirements such as the quality of the acquired images not being too low, and requirements for image resolution, size, etc. In other words, data that does not meet the requirements or is of low quality is excluded.

[0055] Landsat and Sentinel remote sensing data that have been filtered out from the same preset time interval are fused to obtain fused data. The same preset time interval means the same year; in other words, one fused data is obtained for each set of historical remote sensing data.

[0056] For each fused data set, spectral band data is extracted, and vegetation index data is calculated using the extracted spectral band data.

[0057] Continuing with the example above, in this example, Landsat 7 / 8 and Sentinel-2 (LC / S2) surface reflectance (SR) images of the Hulunbuir Grassland available from 2015 to 2021 were obtained.

[0058] The Landsat and Sentinel remote sensing datasets are generated through four main steps: excluding poor-quality observations, fusing LC / S2 images, generating spectral bands and vegetation indices, and constructing and interpolating time series.

[0059] Specifically, firstly, for each image, low-quality observations from clouds, cloud shadows, snow, and scanline corrector gaps are masked as no data using the CFmask method. Secondly, the spectral bands (SBs) of LC / S2 were fused using ordinary least squares (OLS) regression. This method can generate comparable time series based on different sensors of LC / S2. Subsequently, three vegetation indices were calculated using LC / S2 data: Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Surface Water Index (SDI). NDVI and EVI are closely related to vegetation cover, greening, and yield, and are good indicators for monitoring surface water dynamics. The vegetation indices were calculated using surface reflectance values ​​from the blue, red, near-infrared, and short-wavelength infrared bands of LC / S2 images according to equations (1-3). Finally, due to the uneven observation frequency of individual pixels in the original LC / S2 time series dataset, the time series of SBs and VIs were reconstructed at 10-day intervals. NDVI and EVI were synthesized by calculating the maximum values ​​over 10 days, while others were synthesized by calculating the average values ​​over 10 days. If high-quality observation data was unavailable within a 10-day period, data gaps were filled using linear interpolation.

[0060]

[0061] in, Normalized Difference Vegetation Index (NDVI) The surface reflectance is in the near-infrared band. This represents the surface reflectance value in the red light band. To enhance the vegetation index, The surface reflectance is in the blue light band. It is a surface water index. This refers to the surface reflectance in the short-wavelength infrared band.

[0062] After acquiring historical remote sensing data of the measured area, the process proceeds to step 204, where the acquired historical remote sensing data sets are used to determine the characteristic data of the remote sensing data. Specifically, Historical remote sensing data from the same year are divided into a single remote sensing dataset. For each dataset, the average value of each indicator is calculated and used as the feature data for that year. For example, for historical remote sensing data from the same year (e.g., 2009), the average value of the red band surface reflectance is calculated and used as the feature data for red band surface reflectance. Using the same method, a total of eight feature data are obtained: average surface reflectance in red band, green band, blue band, near-infrared band, short-wavelength infrared band, normalized difference vegetation index, enhanced vegetation index, and surface water index.

[0063] Next, in step 206, the feature data is input into the grazing probability prediction model for processing in order to predict the grazing probability of the tested area.

[0064] In some embodiments, the grazing probability prediction model is generated by the following method.

[0065] First, first sample data indicating heavily grazing areas, second sample data indicating ungrazing areas, and historical remote sensing data of the target area are collected within the target area. The process of acquiring historical remote sensing data of the target area can be referred to step 202 above, and will not be repeated here.

[0066] For ease of understanding and description, the first sample data and the second sample data will be collectively referred to as field sample data. Field samples include training samples used to build a grazing probability prediction model and validation samples used to calculate the confusion matrix, and are generally collected by staff in the field.

[0067] For example, staff conducted a field survey of the Geographic Area (GI) in Hulunbuir City in 2021. The field survey covered over 300 sites. For each site, field data was collected, including photographs, grazing management, livestock numbers, vegetation cover, biodiversity, plant height, altitude, latitude, longitude, and surrounding environment. Based on 2021 field photographs, high-resolution images from Google Earth from 2015 to 2021, and the China High Resolution Earth Observation System (CROS), multi-year Regions of Interest (ROIs) for training samples from 2015 to 2021 (60 ROIs) for heavy grazing and 45 ROIs for ungrazing were digitized. The ROIs were drawn according to large-scale pasture shapes, and in terms of pixel count, the training ROIs covered an area of ​​over 714.5 square kilometers with over 7.15 × 10⁶ pixels (10 m × 10 m). Based on field surveys, plant species composition, vegetation and bare soil cover, live and lodged biomass, and manure density, heavily grazing samples were visually identified. Enclosed plots were classified as ungrazing plots. The heavily grazing and ungrazing samples represent two distinct grassland conditions, categorized according to grazing gradients.

[0068] Then, training samples are constructed using historical remote sensing data of the target area, and label data are determined using the first and second sample data. The data in the training samples are the feature data (or the mean of each set of historical remote sensing data) of each group, as detailed in step 204 above, which will not be repeated here.

[0069] The tag data is threshold data indicating grazing intensity. In some embodiments, the threshold can be defined by field samples collected from the target area.

[0070] In a specific example, to determine thresholds for different grazing intensities, researchers utilized the frequency distributions of grazing probabilities from heavily grazing and ungrazing samples. Most of the information was extracted without introducing excessive noise, and thresholds were determined based on cumulative frequencies. In this example, by analyzing field sample data from the Hulunbuir Grassland, a threshold of 0.6 (≥0.6) was obtained for heavily grazing areas, and a threshold of 0.2 (≤0.2) was used for ungrazing areas. Plots were then created using thresholds of 0.6 and 0.2. Figure 3 , Figure 3 A schematic diagram illustrating the grazing probability analysis of an ungrazed sample (a) and a heavily grazed sample (b) according to an embodiment of the present invention is shown. Figure 3 The middle line displays the cumulative frequency in the ungrazed tag data and the heavily grazed tag data. For example... Figure 3As shown, if the grazing probability in the tested area is not less than 0.6, the grazing intensity of the tested area is determined to be heavy grazing. If the grazing probability in the tested area is less than 0.6 but greater than 0.2, the grazing intensity of the tested area is determined to be moderate grazing. If the grazing probability in the tested area is not greater than 0.2, the grazing intensity of the tested area is determined to be light grazing.

[0071] Finally, based on the training set, the initial grazing probability prediction model is trained to obtain a grazing probability prediction model. In this embodiment, the initial grazing probability prediction model is a random forest (RF) model. RF is an ensemble learning algorithm that is more accurate and robust to noise than a single algorithm. The RF model is built using the scikit-learn library in Python, and the model calculates the average probability of all trees in each class as the classification probability.

[0072] The initial grazing probability prediction model is trained using sample data, and the training results are compared using labeled data in order to adjust the parameters and fit the initial grazing probability prediction model.

[0073] In some embodiments, the accuracy of the generated grazing probability prediction model can be verified using a ten-fold cross-validation method.

[0074] Next, in step 208, the grazing intensity of the tested area is determined based on the grazing probability. Specifically, if the grazing probability is not less than a first threshold, the grazing intensity of the tested area is determined to be heavy grazing. If the grazing probability is less than the first threshold but greater than a second threshold, the grazing intensity of the tested area is determined to be moderate grazing. If the grazing probability is not greater than the second threshold, the grazing intensity of the tested area is determined to be light grazing.

[0075] In some embodiments, the determined grazing intensity can be verified by establishing a confusion matrix.

[0076] The method provided by this invention obtains the grazing probability of the measured area through historical remote sensing data of the measured area, and then obtains the grazing intensity through the grazing probability, realizing high-resolution, long-time-series quantitative monitoring of grazing intensity in a large area of ​​the measured area, with flexible extrapolation and high accuracy.

[0077] Figure 4A schematic diagram of a device 400 for determining grazing intensity according to an embodiment of the present invention is shown. The device 400 resides in a computing device. The device 400 includes an acquisition module 402, a first determination module 404, an input module 406, and a second determination module 408, which are coupled to each other. The acquisition module 402 is adapted to acquire at least one set of historical remote sensing data of permanent grassland within a measured area during the hot growing season. The first determination module 404 is adapted to determine characteristic data of the acquired historical remote sensing data using the acquired sets of historical remote sensing data. The input module 406 is adapted to input the characteristic data into a grazing probability prediction model for processing to predict the grazing probability of the measured area. The second determination module 408 is adapted to determine the grazing intensity of the measured area based on the grazing probability.

[0078] It should be noted that the working principle and process of the device 400 provided in this embodiment are similar to those of the method 200 described above. For relevant details, please refer to the description of the method 200 described above, which will not be repeated here.

[0079] The various techniques described herein can be implemented in combination with hardware or software, or a combination thereof. Thus, the methods and apparatus of the present invention, or certain aspects or portions thereof, can take the form of program code (i.e., instructions) embedded in a tangible medium, such as a removable hard disk, USB flash drive, floppy disk, CD-ROM, or any other machine-readable storage medium, wherein when the program is loaded into and executed by a machine such as a computer, the machine becomes an apparatus for practicing the present invention.

[0080] When the program code is executed on a programmable computer, the computing device generally includes a processor, a processor-readable storage medium (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device. The memory is configured to store program code; the processor is configured to execute the method of the present invention according to instructions in the program code stored in the memory.

[0081] A8. The method as described in A3, wherein the first remote sensing data is MODIS-MCD12Q1 remote sensing data; and determining the permanent grassland area included in the measured area based on the first remote sensing data of the measured area, including: extracting grassland maps within a preset time period from the MODIS-MCD12Q1 remote sensing data; and taking the areas marked as grassland in the grassland maps within the preset time period as the permanent grassland area. A9. The method as described in A3, wherein the second remote sensing data is MODIS-MOD11A2 remote sensing data; and determining the thermal growth season cycle of the measured area based on the second remote sensing data of the measured area, including: extracting nighttime temperature time series data from the MODIS-MOD11A2 remote sensing data; determining the first date when the first nighttime minimum temperature is less than 5°C and the second date when the last nighttime maximum temperature is greater than 5°C from the nighttime temperature time series data, and taking the time period between the first date and the second date as the thermal growth season cycle. A10. The method described in A2 further includes a step of generating the grazing probability prediction model: collecting first sample data indicating a heavily grazing area, second sample data indicating an ungrazing area, and historical remote sensing data of the target area within the target area; constructing training samples using the historical remote sensing data of the target area, and determining label data using the first and second sample data; and training the initial grazing probability prediction model based on the training samples and label data to obtain the grazing probability prediction model.

[0082] By way of example, and not limitation, readable media include readable storage media and communication media. Readable storage media stores information such as computer-readable instructions, data structures, program modules, or other data. Communication media generally embodies computer-readable instructions, data structures, program modules, or other data in the form of modulated data signals such as carrier waves or other transmission mechanisms, and includes any information delivery medium. Any combination of the above is also included within the scope of readable media.

[0083] In the specification provided herein, the algorithms and displays are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used with the examples of this invention. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing preferred embodiments of the invention.

[0084] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0085] Similarly, it should be understood that, in order to streamline this disclosure and aid in understanding one or more of the various aspects of the invention, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof. However, this method of disclosure should not be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the following claims, inventive aspects lie in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into this detailed description, wherein each claim itself is a separate embodiment of the invention.

[0086] Those skilled in the art will understand that modules, units, or components of the devices disclosed in the examples herein can be arranged in the devices described in this embodiment, or alternatively, can be located in one or more devices different from the devices in this example. The modules in the foregoing examples can be combined into a single module or, in addition, can be divided into multiple sub-modules.

[0087] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0088] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features but not others included in other embodiments, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0089] Furthermore, some of the embodiments described herein are methods or combinations of method elements that can be implemented by a processor of a computer system or by other means of performing the functions. Therefore, a processor having the necessary instructions for implementing the methods or method elements forms means for implementing the methods or method elements. Furthermore, the elements described herein in the apparatus embodiments are examples of means for implementing the functions performed by elements for the purposes of carrying out the invention.

[0090] As used herein, unless otherwise specified, the use of ordinal numbers such as “first,” “second,” “third,” etc., to describe ordinary objects merely indicates different instances of similar objects and is not intended to imply that the objects being described must have a given order in time, space, ordering, or any other manner.

[0091] Although the invention has been described with respect to a limited number of embodiments, those skilled in the art will understand from the foregoing description that other embodiments are conceivable within the scope of the invention described herein. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and edibility purposes, and not for the purpose of interpreting or limiting the subject matter of the invention. Therefore, many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims. The disclosure of the invention is illustrative rather than restrictive, and the scope of the invention is defined by the appended claims.

Claims

1. A method for determining grazing intensity, performed in a computing device, the method comprising: Acquire at least one set of historical remote sensing data on permanent grasslands within the surveyed area during the hot growing season; Using the acquired historical remote sensing data, the characteristic data of the remote sensing data are determined; The feature data is input into the grazing probability prediction model for processing in order to predict the grazing probability of the tested area. Based on the grazing probability, the grazing intensity of the tested area is determined; The remote sensing data includes at least spectral band data and vegetation index data. At least one set of historical remote sensing data for permanent grassland within the measured area during the hot growing season is acquired, including: determining the permanent grassland area included in the measured area based on the first remote sensing data of the measured area; determining the hot growing season cycle of the measured area based on the second remote sensing data of the measured area; acquiring Landsat and Sentinel remote sensing data corresponding to the permanent grassland area within the same hot growing season cycle; and using the acquired Landsat and Sentinel remote sensing data to determine the spectral band data and vegetation index data. The first remote sensing data is MODIS-MCD12Q1 remote sensing data, and based on the first remote sensing data of the measured area, the permanent grassland area included in the measured area is determined, including: extracting grassland maps within a preset time period from the MODIS-MCD12Q1 remote sensing data, and taking the areas marked as grassland in the grassland maps within the preset time period as the permanent grassland area. The second remote sensing data is MODIS-MOD11A2 remote sensing data, and based on the second remote sensing data of the measured area, the thermal growth season cycle of the measured area is determined, including: extracting nighttime temperature time series data from the MODIS-MOD11A2 remote sensing data, determining the first date when the first nighttime minimum temperature is less than 5°C and the second date when the last nighttime maximum temperature is greater than 5°C from the nighttime temperature time series data, and taking the time period between the first date and the second date as the thermal growth season cycle; The steps for generating the grazing probability prediction model are as follows: Collect first sample data of areas indicating heavy grazing intensity, second sample data of areas indicating ungrazing intensity, and historical remote sensing data of the target area within the target area; Using historical remote sensing data of the target area, training samples are constructed, and label data is determined using the first sample data and the second sample data. Based on the training samples and label data, the initial grazing probability prediction model is trained to obtain the grazing probability prediction model.

2. The method as described in claim 1, wherein, Based on the grazing probability, the grazing intensity of the tested area is determined, including: If the grazing probability is not less than the first threshold, then the grazing intensity of the tested area is determined to be heavy grazing. If the grazing probability is less than the first threshold and greater than the second threshold, then the grazing intensity of the tested area is determined to be moderate grazing. If the grazing probability is not greater than the second threshold, then the grazing intensity of the tested area is determined to be light grazing.

3. The method as described in claim 1, wherein, The spectral data includes at least one of the following indicators: surface reflectance in the red, green, blue, near-infrared, and short-wavelength infrared bands; and The vegetation index data includes at least one of the following indicators: normalized vegetation index, enhanced vegetation index, and surface water index.

4. The method of claim 3, wherein, Acquire Landsat and Sentinel remote sensing data corresponding to the permanent grassland area, and use the acquired Landsat and Sentinel remote sensing data to determine the spectral band data and vegetation index data, including: The Cfmask method was used to filter out data in Landsat and Sentinel remote sensing data that did not meet the preset conditions. The Landsat and Sentinel remote sensing data that were filtered out within the same preset time interval were fused to obtain the fused data. For each fused data set, spectral band data is extracted from the fused data set, and the vegetation index data is calculated using the extracted spectral band data.

5. The method of claim 4, wherein, The vegetation index data for each fused dataset is calculated using the following formula: , in, Normalized Difference Vegetation Index (NDVI) The surface reflectance is in the near-infrared band. This represents the surface reflectance value in the red light band. To enhance the vegetation index, The surface reflectance is in the blue light band. This is a surface water index. This refers to the surface reflectance in the short-wavelength infrared band.

6. The method of claim 3, wherein, Using the acquired historical remote sensing data, the characteristic data of the remote sensing data are determined, including: Historical remote sensing data from the same year are divided into a single remote sensing dataset; For each remote sensing dataset, the average value of each indicator in the dataset is calculated and used as the feature data for the corresponding year.

7. A device for determining grazing intensity, residing in a computing device, the device comprising: The acquisition module is suitable for acquiring at least one set of historical remote sensing data of permanent grassland in the measured area during the hot growing season; The first determining module is adapted to use the acquired sets of historical remote sensing data to determine the feature data of the remote sensing data; The input module is adapted to input the feature data into the grazing probability prediction model for processing, so as to predict the grazing probability of the tested area. The second determining module is adapted to determine the grazing intensity of the tested area based on the grazing probability; The remote sensing data includes at least spectral band data and vegetation index data. At least one set of historical remote sensing data for permanent grassland within the measured area during the hot growing season is acquired, including: determining the permanent grassland area included in the measured area based on the first remote sensing data of the measured area; determining the hot growing season cycle of the measured area based on the second remote sensing data of the measured area; acquiring Landsat and Sentinel remote sensing data corresponding to the permanent grassland area within the same hot growing season cycle; and using the acquired Landsat and Sentinel remote sensing data to determine the spectral band data and vegetation index data. The first remote sensing data is MODIS-MCD12Q1 remote sensing data, and based on the first remote sensing data of the measured area, the permanent grassland area included in the measured area is determined, including: extracting grassland maps within a preset time period from the MODIS-MCD12Q1 remote sensing data, and taking the areas marked as grassland in the grassland maps within the preset time period as the permanent grassland area. The second remote sensing data is MODIS-MOD11A2 remote sensing data, and based on the second remote sensing data of the measured area, the thermal growth season cycle of the measured area is determined, including: extracting nighttime temperature time series data from the MODIS-MOD11A2 remote sensing data, determining the first date when the first nighttime minimum temperature is less than 5°C and the second date when the last nighttime maximum temperature is greater than 5°C from the nighttime temperature time series data, and taking the time period between the first date and the second date as the thermal growth season cycle; The steps for generating the grazing probability prediction model are as follows: Collect first sample data of areas indicating heavy grazing intensity, second sample data of areas indicating ungrazing intensity, and historical remote sensing data of the target area within the target area; Using historical remote sensing data of the target area, training samples are constructed, and label data is determined using the first sample data and the second sample data. Based on the training samples and label data, the initial grazing probability prediction model is trained to obtain the grazing probability prediction model.

8. A computing device, comprising: At least one processor; and A memory storing program instructions configured to be executed by the at least one processor, the program instructions including instructions for performing the method as described in any one of claims 1-6.

9. A readable storage medium storing program instructions that, when read and executed by a computing device, cause the computing device to perform the method as described in any one of claims 1-6.

Images