Integrated sky-ground monitoring method for biomass of typical vegetation in water-level-fluctuating zone of three gorges reservoir

Abstract

The application discloses a kind of Three Gorges Reservoir belt typical vegetation biomass sky-ground integrated monitoring method, the method utilizes satellite remote sensing (space-based), unmanned aerial vehicle aerial photography (air-based), ground measurement (ground-based) three combination means, with normalized difference vegetation index NDVI as core, respectively in May, June, July, August four time periods to the study area sky-ground vegetation index extraction, observation, finally establish low-middle-high resolution based on satellite NDVI, unmanned aerial vehicle NDVI, ground measurement data's belt typical vegetation biomass inversion model, realize belt in the above-ground fresh weight of Xanthium, dog tooth grass, mixed type vegetation biomass, air-based unmanned aerial vehicle NDVI data, and the conversion between space-based sentinel NDVI data and air-based unmanned aerial vehicle NDVI remote sensing data, so as to improve the precision of biomass inversion, provide reference basis for Three Gorges Reservoir belt vegetation recovery effect evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of ecological environment monitoring and evaluation, and particularly relates to a method for monitoring the biomass of typical vegetation in the Three Gorges Reservoir area by integrating sky and ground. BACKGROUND

[0002] The drawdown zone is also known as the drawdown area, which is a unique phenomenon of rivers, lakes and reservoirs. It is mainly formed by two reasons. One is the seasonal water level fluctuation, which mainly refers to the area of the submerged land periodically exposed to the surface of the water due to the seasonal water level fluctuation. In addition, the drawdown zone caused by special climate (such as the water level drop of Dongting Lake due to drought) is also included. The other is the periodic water storage, which mainly refers to the water level fluctuation caused by the periodic flood storage or discharge of large reservoirs (such as the Three Gorges Dam). The research object of the present project is the latter.

[0003] After the completion of the Three Gorges Project, the water level for power generation in winter is 175 meters, and the water level for flood control in summer is reduced to 145 meters. The land exposed by the 30-meter water level difference is the drawdown zone. The main stream of the Yangtze River is the flood season from June to September every year, and the dry season is from the end of September to the next May. The Three Gorges Reservoir starts to store water at the end of September or the beginning of October when the flood season is about to end, and the water level reaches the designed normal water storage level of 175 meters. The water level of the drawdown zone rises and falls against the natural flood and drought law, and the exposed land is the hottest and most humid during the period of exposure to land. Therefore, the Three Gorges Reservoir has the least water storage in summer, the largest exposed area of the drawdown zone, and the most abundant vegetation. The research period is selected from April to September.

[0004] The "sky-ground" integrated ecological environment monitoring system will provide strong technical support for ecological environment monitoring and become a new direction for the future development of environmental monitoring.

[0005] The "sky-ground" integrated ecological environment monitoring is a three-dimensional ecological environment monitoring sensing system that comprehensively uses satellite remote sensing monitoring, aerial remote sensing monitoring and ground fixed monitoring, and is based on key technologies such as data mining, data fusion, data collaboration and data assimilation to obtain more accurate data support. The "sky-ground" integrated ecological environment monitoring system can more comprehensively reflect the current situation and development trend of environmental quality, and can comprehensively and finely evaluate the ecological environment status before and after the implementation of major ecological restoration projects, and provide scientific basis for macro environmental management and environmental planning.

[0006] At present, the construction of the integrated ecological environment monitoring system of sky-ground is still in its infancy in China, and there are still problems such as high construction cost, non-uniform data fusion standard, and urgent need to establish data sharing mechanism. The integrated construction of sky-ground needs to coordinate satellite remote sensing, ground stations and monitoring data obtained by aviation, realize data fusion and data sharing, and truly build a three-dimensional monitoring system of satellite, aviation and ground monitoring. In recent years, the near-earth (50-100m) remote sensing information acquisition technology of developing unmanned aerial vehicle as a remote sensing platform has become an ideal link to build the integrated remote sensing system of sky-ground. At the same time, with the popularity of digital cameras, photography technology has become a hot digital image information acquisition method. The camera based on digital image technology of plant biomass, leaf area index (LAI) and normalized vegetation index band is specially designed for measurement purpose. Therefore, an integrated sky-ground monitoring method for dynamically monitoring the biomass of typical vegetation in the drawdown zone of the Three Gorges Reservoir Area is urgently needed. SUMMARY

[0007] Therefore, the purpose of the present application is to provide a sky-ground integrated monitoring method for the biomass of typical vegetation in the drawdown zone of the Three Gorges Reservoir Area.

[0008] In order to achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0009] 1. A sky-ground integrated monitoring method for the biomass of typical vegetation in the drawdown zone of the Three Gorges Reservoir Area, comprising the following steps:

[0010] According to the principles of uniform distribution, consideration of topography, good flight conditions and diverse land cover types, and taking into account the topography of the drawdown zone, the vegetation types inside the drawdown zone, the external conditions and location factors of the drawdown zone, a representative, universal, operable and typical area of the drawdown zone in the Three Gorges Reservoir Area is selected as the research area. Satellite remote sensing, unmanned aerial vehicle aerial photography and ground measurement are combined, and the normalized vegetation index NDVI is taken as the core. The sky-ground-vegetation index of the research area is extracted and observed in four time periods of May, June, July and August. Finally, a low-medium-high resolution drawdown zone typical vegetation biomass inversion model based on satellite NDVI, unmanned aerial vehicle NDVI and ground measurement data is established, and the spatio-temporal dynamic change characteristics of the biomass of the drawdown zone during the vegetation growing season from May to August are analyzed.

[0011] The satellite remote sensing adopts the European Sentinel-2 satellite; the unmanned aerial vehicle aerial photography adopts the DJI Spark 4 multi-spectral version unmanned aerial vehicle; and the ground measurement adopts ground sample or sample line investigation, and the investigation content is the name, land use type, height, biomass and distribution position of the ground vegetation, and the ground NDVI is measured by using the field crop NDVI measuring instrument.

[0012] In the ground plot or line investigation, in order to improve the inversion precision of the typical vegetation biomass of the drawdown zone, the ground measured data is divided into three groups of Cynodon, Xanthium and mixed for analysis, wherein the mixed refers to the remaining vegetation types in the drawdown zone except Cynodon and Xanthium.

[0013] The biomass is aboveground fresh weight, underground dry weight, underground fresh weight, underground dry weight, and total fresh weight and total dry weight.

[0014] Preferably, the drawdown zone terrain is one or more of steep slope type, flat type, gentle slope type and reservoir tail type; the drawdown zone internal vegetation type is one or more of Xanthium, Cynodon, Eragrostis poii, Hemarthria altissima, Cyperus rotundus, Scopamphium brevifolium and Ambrosia artemisiifolia; and the drawdown zone external condition is one or more of land use type, including arable land, forest land, shrub land, orchard, garden land, grassland, bare land, water area and residential area.

[0015] Preferably, the research area is five regions of Xueshangtangwan in Wushan, Dalangba in Kai County, Shipangou in Yunyang, Hongyanzi in Zhongxian and Muhecun in Fuling, covering two tributary points and three main stream points of the Three Gorges Reservoir area.

[0016] Preferably, in the ground monitoring, in order to ensure data quality, the ground NDVI monitoring points and plot points of each research area are not less than 5, the line is not less than 1, and 10 point positions are ensured in each research area 1km 2 , 5 groups of values are measured in the range of 1Mx 1M, and the average value is taken.

[0017] Preferably, the drawdown zone typical vegetation biomass inversion model comprises an inversion model between the biomass data obtained by ground measurement and the UAV NDVI data, and an inversion model between the Sentinel NDVI and the UAV NDVI.

[0018] Preferably, the construction method of the inversion model between the biomass data obtained by ground measurement and the UAV NDVI data is as follows: the UAV NDVI data corresponding to the time and space is obtained according to the ground monitoring point, and then the aboveground fresh weight, aboveground dry weight, underground fresh weight, underground dry weight, total fresh weight and total dry weight of Xanthium, Cynodon and mixed plot are respectively analyzed to obtain the relationship between the biomass of different vegetation types and NDVI, and finally the inversion model between the biomass data obtained by ground measurement and the UAV NDVI data is constructed by comprehensively considering the biological significance and the rationality of correlation.

[0019] Further preferably, in the inversion model between the biomass data obtained by ground measurement and the UAV NDVI data, the fitting equation of the aboveground fresh weight (Y) of Xanthium biomass and the UAV NDVI (X) is as follows:

[0020] Y = 224062X2 -331611X + 124488 (R 2 = 0.7368);

[0021] The fitting equation of the aboveground fresh weight of Bidens pilosa (Y) and UAV NDVI (X) is:

[0022] Y = 166267X 2 -244406X + 90193 (R 2 = 0.7161);

[0023] The fitting equation of the total dry weight of Bidens pilosa (Y) and UAV NDVI (X) is:

[0024] Y = 174560X 2 -256952X + 95166 (R 2 = 0.7246);

[0025] The fitting equation of the aboveground fresh weight of Cynodon dactylon (Y) and UAV NDVI (X) is:

[0026] Y = 30940X 2 -36279X + 11316 (R 2 = 0.5515);

[0027] The fitting equation of the aboveground dry weight of Cynodon dactylon (Y) and UAV NDVI (X) is:

[0028] Y = 24728X 2 -28505X + 8454.4 (R 2 = 0.6548);

[0029] The fitting equation of the total dry weight of Cynodon dactylon (Y) and UAV NDVI (X) is:

[0030] Y = 20800X 2 -23020X + 6893.1 (R 2 = 0.6264);

[0031] The fitting equation of the aboveground fresh weight of mixed plants (Y) and UAV NDVI (X) is:

[0032] Y = 13694X 2 -15468X + 5525 (R 2 = 0.5254).

[0033] Preferably, the method for constructing the inversion model between the sentinel NDVI and the UAV NDVI is as follows: taking the sentinel data of different months in the study area as a reference, the corresponding month and the same place of the UAV image are selected for comparison, and the same points under different land use types such as full-coverage vegetation, bare land, water area and house are synchronously selected; since the spatial resolution of the sentinel data is 10 m and the spatial resolution of the UAV is 0.1 m, the average value of the NDVI in the range of 100*100 pixels of the UAV image is calculated to match the sentinel data.

[0034] Further preferably, in the inversion model between the sentinel NDVI and the UAV NDVI, the fitting equation of the sentinel NDVI (X) and the UAV NDVI (Y) is as follows:

[0035] Y = 1.3695X-0.3007 (R 2 = 0.954).

[0036] The method for monitoring the biomass of the typical vegetation in the Three Gorges Reservoir belt disclosed by the application is combined with satellite remote sensing (space-based), UAV aerial photography (air-based) and ground measurement (ground-based), takes the normalized vegetation index NDVI as the core, extracts and observes the space-air-ground vegetation index of the study area in four time periods of May, June, July and August, finally establishes a low-medium-high resolution biomass inversion model of the typical vegetation in the Three Gorges Reservoir belt based on the satellite NDVI, the UAV NDVI and the ground measurement data, analyzes the spatial and temporal dynamic change characteristics of the biomass of the vegetation in the Three Gorges Reservoir belt during the vegetation growth season from May to August, realizes the conversion between the aboveground fresh weight of the biomass of the Xanthium sibiricum, the Cynodon dactylon and the mixed type vegetation, the air-based UAV NDVI data and the space-based sentinel NDVI data and the air-based UAV NDVI remote sensing data, thereby improving the biomass inversion precision and providing a reference basis for the evaluation of the vegetation restoration effect in the Three Gorges Reservoir belt. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to make the purpose, technical scheme and beneficial effects of the application more clear, the application provides the following drawings for description:

[0038] Figure 1 It is a distribution map of the UAV and the ground survey point;

[0039] Figure 2 It is a technical roadmap of the application;

[0040] Figure 3 It is a European Sentinel-2 remote sensing image data;

[0041] Figure 4 It is an appearance of the DJI sprite 4 multispectral version UAV.

[0042] Figure 5 For UAV aerial photography data results (part);

[0043] Figure 6 For field crop NDVI measuring instrument CropSense;

[0044] Figure 7 For NDVI measured point position data (part);

[0045] Figure 8 For Xanthium ground fresh weight and UAV NDVI fitting curve;

[0046] Figure 9 For Xanthium ground dry weight and UAV NDVI fitting relationship;

[0047] Figure 10 For Xanthium underground fresh weight and UAV NDVI fitting curve;

[0048] Figure 11 For Xanthium underground dry weight and UAV NDVI fitting relationship;

[0049] Figure 12 For Xanthium total fresh weight and UAV NDVI fitting curve;

[0050] Figure 13 For Xanthium total dry weight and UAV NDVI fitting curve;

[0051] Figure 14 For Bermudagrass ground fresh weight and UAV NDVI fitting curve;

[0052] Figure 15 For Bermudagrass ground dry weight and UAV NDVI fitting curve;

[0053] Figure 16 For Bermudagrass underground fresh weight and UAV NDVI fitting curve;

[0054] Figure 17 For Bermudagrass underground dry weight and UAV NDVI fitting curve;

[0055] Figure 18 For Bermudagrass total fresh weight and UAV NDVI fitting curve;

[0056] Figure 19 For Bermudagrass total dry weight and UAV NDVI fitting curve;

[0057] Figure 20 For mixed ground fresh weight and UAV NDVI fitting curve;

[0058] Figure 21 For mixed ground dry weight and UAV NDVI fitting curve;

[0059] Figure 22 Fitting curve of mixed type underground fresh weight and UAV NDVI;

[0060] Figure 23 Fitting curve of mixed type underground dry weight and UAV NDVI;

[0061] Figure 24 Fitting curve of mixed type total fresh weight and UAV NDVI;

[0062] Figure 25 Fitting curve of mixed type total dry weight and UAV NDVI;

[0063] Figure 26 Fitting relationship diagram of Fuling August sentinel NDVI and UAV NDVI;

[0064] Figure 27 Fitting relationship diagram of Zhongxian July sentinel NDVI and UAV NDVI;

[0065] Figure 28 Fitting relationship diagram of Wushan June sentinel NDVI and UAV NDVI;

[0066] Figure 29 Fitting relationship diagram of three places sentinel NDVI and UAV NDVI;

[0067] Figure 30 Diagram of UAV image and actual positioning error. DETAILED DESCRIPTION

[0068] The present application will be further described below in conjunction with the drawings and specific examples, so that those skilled in the art can better understand the present application and implement it, but the examples are not as a limitation on the present application.

[0069] Example 1, determine the research area, research content and research method

[0070] I. Research area selection principle

[0071] (1) Uniform distribution

[0072] The research area selection should cover the main stream of the Yangtze River, tributaries, and areas south of the Yangtze River, north of the Yangtze River, and be evenly distributed along the water-level fluctuation zone in Chongqing, to ensure that the research area selection is representative.

[0073] (2) Topography

[0074] The water-level fluctuation zone is divided into steep slope type, flat type, island type, reservoir tail type and canyon type, and all types should be covered as much as possible when selecting the research area, to ensure that the research area selection is universal.

[0075] (3) Flight conditions are good

[0076] The UAV flight should be chosen as relatively flat, open area, avoid the airport, military facilities and other no-fly zones and sensitive areas such as railway, high-voltage line tower, and at the same time, there are no high-rise buildings, mountains and other obstacles that can block the signal in the flight range, to ensure that the study area has operability.

[0077] (4) Land cover type is diverse

[0078] Affected by climate conditions, topography, the inner zone is mainly composed of herbaceous and shrub vegetation such as dogtooth grass and goldenrod, and there are farmland, woodland, garden and a small amount of residents outside the zone. The study area should consider the internal and external conditions of the zone to ensure that the study area is typical.

[0079] Based on the above principles, this study selected Wushan Xuetang Bay, Kaizhou Dalangba, Yunyang Shibangou, Zhongxian Hongyanzi, and Fuling Muhecun five areas, including two tributary points and three Yangtze River mainstream points. The UAV and ground survey point distribution map is shown in Figure 1 , and the study area condition statistics are shown in Table 1.

[0080] Table 1 Statistics of study area

[0081]

[0082] II. Research content and goal

[0083] This study mainly focuses on the seasonal variation characteristics and overall distribution of natural / artificial restoration of plants in the zone from May to August when the water level in the Three Gorges Reservoir area drops to 145m. Satellite remote sensing, UAV aerial photography, and ground monitoring are used to establish a zone vegetation monitoring system. The monitoring content includes microbial quantity, vegetation index, etc.

[0084] The selection principle of ground cover mainly studies the ground land use type, patch number, area, diversity, spatial distribution law, etc. The vegetation index mainly includes NDVI index.

[0085] III. Research methods and technical route

[0086] This study mainly takes NDVI vegetation coverage index as the core, and extracts and observes the vegetation index in the sky-ground in four time periods of May, June, July and August in 2021 to build a zone vegetation sky-ground integrated monitoring system. The technical route of the invention is shown in Figure 2 .

[0087] 1. Satellite remote sensing monitoring

[0088] "Sky" refers to large-scale environmental monitoring of the ecological environment and its changing trends using satellite remote sensing technology. Satellite remote sensing environmental monitoring uses artificial satellites such as Gaofen-1, Gaofen-2, Gaofen-6, Ziyuan-3, and Huanjing-1 as platforms. It utilizes visible light, infrared, and microwave detection instruments to conduct remote sensing monitoring of large-scale macroscopic environmental quality and ecological conditions through photography or scanning, information sensing, transmission, and processing. The accuracy can reach the meter level, enabling multi-time-series, large-scale monitoring of the study area.

[0089] This study primarily utilizes remote sensing image data from the European Sentinel-2 satellite, such as... Figure 3 As shown. Sentinel-2 is a high-resolution multispectral imaging satellite carrying a multispectral imager (MSI) for land monitoring. It provides images of vegetation, soil and water cover, inland waterways, and coastal areas, and can also be used for emergency rescue services. It consists of two satellites, 2A and 2B. Each satellite carries a multispectral imager (MSI) at an altitude of 786 km, covering 13 spectral bands with a swath width of 290 km. The ground resolutions are 10 m, 20 m, and 60 m. The revisit period for a single satellite is 10 days, while the two complementary satellites have a revisit period of 5 days. It offers varying spatial resolutions from visible and near-infrared to short-wave infrared. In optical data, Sentinel-2 data is unique in containing three bands within the red-edge range, making it highly effective for monitoring vegetation health.

[0090] The drawdown zone is narrow, so Sentinel imagery, with its high temporal and spatial resolution, is ideal data for remote sensing monitoring of the drawdown zone in the Three Gorges Reservoir area.

[0091] 2. Aerial remote sensing monitoring

[0092] "Airborne" refers to the use of airborne remote sensing technology to monitor the environmental status of key areas of concern at small to medium scales. Airborne remote sensing environmental monitoring utilizes various aircraft, airships, and high-altitude balloons as sensor carriers, equipped with specific sensors to monitor individual environmental events and local or regional environmental conditions. Compared to satellite remote sensing, UAV monitoring is less expensive, simpler to operate, and offers centimeter-level accuracy. It can conduct low-altitude monitoring under harsh conditions, has a short data acquisition cycle, and high timeliness, making it suitable for monitoring landscape types, number of patches, area, density, diversity, and spatial distribution patterns in small-scale drawdown zones.

[0093] This study utilized a small drone (DJI Phantom 4 Multispectral Edition, equipped with six 1 / 2.9-inch CMOS sensors, including one color sensor for visible light imaging and five monochrome sensors for multispectral imaging; each sensor has 2.08 million effective pixels, totaling 2.12 million pixels, with a maximum image resolution of 1600×1300) to acquire low-altitude (height: 600FT / 182.88m, accuracy: 9.7cm) images and NDVI data. The DJI Phantom 4 Multispectral Edition drone... Figure 4 As shown, some of the drone aerial photography data results are as follows: Figure 5 As shown. ENVI software was used to interpret aerial imagery data, yielding centimeter-level land use types for five study areas. Simultaneously, combined with high temporal resolution Sentinel-2 data, a model for vegetation cover and biomass inversion in the drawdown zone was established based on quadrat measured data and NDVI vegetation index data. This model was used to analyze the spatiotemporal dynamics of the drawdown zone during the 2021 growing season (May-August).

[0094] 3. Ground monitoring

[0095] "Ground" refers to the precise and dynamic monitoring of specific items such as vegetation, soil, water quality, air quality, and industrial pollution discharge within a designated area using manual laboratory monitoring or online automated monitoring. Combining handheld instruments and ground quadrat surveys, it enables species-level monitoring of vegetation in the drawdown zone, including chlorophyll, NDVI (Normalized Difference Vegetation Index), LAI (Leaf Area Index), biomass, name, type, height, cover, and distribution location.

[0096] Ground quadrat / transect surveys were conducted in five study areas, including the name, type, height, biomass, and distribution of ground vegetation. The ground vegetation index monitoring equipment used was the CropSense, a novel field crop NDVI measuring instrument developed by the Beijing Agricultural Information Technology Research Center. Figure 6 As shown in the figure, it can acquire key growth indicators of vegetation in real time, such as NDVI, leaf area index (LAI), estimated yield, recommended nitrogen, cover, and chlorophyll content. To ensure data quality, each area should have no fewer than 5 ground NDVI monitoring points and quadrat points (with no fewer than 1 transect).

[0097] The ground survey of the drawdown zone was conducted when the vegetation had fully grown out. Specific survey indicators are shown in Table 2.

[0098] Table 2. Sample Form for Ground Survey of Vegetation in the Drawdown Zone

[0099]

[0100] Note: Land types include cultivated land, forest land, orchard land, grassland, bare land, water area, and construction land.

[0101] Ground NDVI measurements ensure 1km 2 No less than 10 points, each point in 1M*1M range to measure 5 groups of values to take the average value, to ensure the quality of the measured data, vegetation NDVI ground measurement table (sample table) as shown in Table 3.

[0102] Table 3 Vegetation NDVI ground measurement table (sample table)

[0103]

[0104] NDVI measured point data (part) as Figure 7 shown.

[0105] Finally, a low-mid-high resolution vegetation biomass inversion model of the hydro-fluctuation belt is established based on Sentinel NDVI, UAV NDVI, and measured data, and the spatiotemporal dynamic change characteristics of the biomass in the hydro-fluctuation zone during the vegetation growing season (May-August) in 2021 are analyzed.

[0106] Example 2, construction of a sky-ground integrated monitoring system

[0107] I. Ground measured data

[0108] The data collection time of this study is from May 13, 2021 to August 25, 2021, and the relevant research areas of 5 counties are surveyed by UAV aerial photography and ground measurement, forming 4 periods of data in May, June, July, and August. Influenced by meteorological factors and instruments, 16 UAV aerial images are obtained, and 73 plant quadrats are harvested.

[0109] Table 4 Statistics of hydro-fluctuation belt quadrat data acquisition

[0110]

[0111]

[0112] The field survey results show that the vegetation types in the hydro-fluctuation belt are mainly Cynodon dactylon and Xanthium strumarium, so the ground measured data is divided into three groups of Cynodon dactylon, Xanthium strumarium, and mixed (Cynodon dactylon and Xanthium strumarium and other vegetation species) for analysis in the construction of the sky-ground integrated monitoring system.

[0113] The measured data of Cynodon dactylon in the hydro-fluctuation belt is 60. The aboveground fresh weight, aboveground dry weight, underground fresh weight, underground dry weight, and total fresh weight and total dry weight of Cynodon dactylon are counted, and the statistical results are shown in Table 5.

[0114] Table 5 Statistics of Cynodon dactylon biomass data in the hydro-fluctuation belt

[0115]

[0116] Table 5 shows that the average value of the dog tooth grass biomass is the largest total fresh weight 3044.49g, the smallest is the underground dry weight 362.62g; the largest standard deviation is the total fresh weight 1992.67g, the smallest is the underground dry weight 265.6g.

[0117] The measured data of the biomass of the Xiang'e in the drawdown zone is 48, and the aboveground fresh weight, underground dry weight, underground fresh weight, underground dry weight, and total fresh weight and total dry weight of the Xiang'e are counted, and the counting results are shown in Table 6.

[0118] Table 6 Biomass data statistics of Xiang'e in drawdown zone

[0119]

[0120] Table 6 shows that the average value of the Xiang'e biomass is the largest total fresh weight 4223.05g, the smallest is the underground dry weight 249.81g; the largest standard deviation is the total fresh weight 3400.45g, the smallest is the underground dry weight 101.94g.

[0121] The measured data of the biomass of the mixed type in the drawdown zone is 168, and the aboveground fresh weight, underground dry weight, underground fresh weight, underground dry weight, and total fresh weight and total dry weight of the mixed type are counted, and the counting results are shown in Table 7.

[0122] Table 7 Biomass data statistics of mixed type

[0123]

[0124] Table 7 shows that the average value of the mixed type biomass is the largest total fresh weight 3480.67g, the smallest is the underground dry weight 311.82g; the largest standard deviation is the total fresh weight 2485.07g, the smallest is the underground dry weight 270.46g.

[0125] II. Data modeling of air base (unmanned aerial vehicle)

[0126] In this study, 44 pieces of unmanned aerial vehicle NDVI data of the drawdown zone are obtained, and the spatial resolution is mainly 0.1m, and the data amount is 11.7G. Among them, 4 pieces of data in Wushan, data amount 0.92G; 10 pieces of data in Yunyang, data amount 2.07G; 12 pieces of data in Fuling, data amount 4G; 12 pieces of data in Kaizhou, data amount 3.23G; 6 pieces of data in Zhongxian, data amount 1.48G.

[0127] On this basis, the series of biomass data obtained by ground measurement and unmanned aerial vehicle NDVI data are analyzed correspondingly to build a sky-ground integrated monitoring system of biomass.

[0128] Specific process is, according to the ground measured point position obtains corresponding time and space on unmanned aerial vehicle NDVI data, again with the ground fresh weight, ground dry weight, underground fresh weight, underground dry weight, total fresh weight and total dry weight of the mixed class sample of Xanthium and Cynodon respectively Linear analysis, explore the relationship between biomass and NDVI of different vegetation types, finally comprehensive consideration biological significance and correlation reasonable construction of the falling zone biomass sky-ground integrated monitoring system.

[0129] 1, Xanthium biomass and unmanned aerial vehicle NDVI analysis

[0130] Xanthium biomass aboveground fresh weight and unmanned aerial vehicle NDVI fitting relationship see Figure 8 Xanthium biomass aboveground fresh weight (Y) and unmanned aerial vehicle NDVI (X) fitting equation is:

[0131] Y=224062X 2 -331611X+124488 (R 2 =0.7368)

[0132] Fitting results show that Xanthium biomass aboveground fresh weight and unmanned aerial vehicle NDVI fitting relationship is good, the correlation coefficient R 2 reached more than 0.7, both showed significant positive correlation.

[0133] Xanthium biomass aboveground dry weight and unmanned aerial vehicle NDVI fitting relationship see Figure 9 Xanthium biomass aboveground dry weight (Y) and unmanned aerial vehicle NDVI (X) fitting equation is:

[0134] Y=166267X 2 -244406X+90193 (R 2 =0.7161)

[0135] Fitting results show that Xanthium biomass aboveground dry weight and unmanned aerial vehicle NDVI fitting relationship is good, the correlation coefficient R 2 reached more than 0.7, both showed significant positive correlation.

[0136] Xanthium biomass underground fresh weight and unmanned aerial vehicle NDVI fitting relationship see Figure 10 Xanthium biomass underground fresh weight (Y) and unmanned aerial vehicle NDVI (X) fitting equation is:

[0137] Y=120917X 2 +186801X-71160 (R 2 =0.2648)

[0138] Fitting results show that Xanthium biomass underground fresh weight and unmanned aerial vehicle NDVI fitting relationship is poor, the correlation coefficient R 2 below 0.3.

[0139] The fitting relationship between the underground dry weight of the biomass of Xanthium sibiricum and the UAV NDVI is shown in Figure 11 The fitting equation between the underground dry weight of the biomass of Xanthium sibiricum (Y) and the UAV NDVI (X) is:

[0140] Y = 8293X 2 -12546X + 4973.1 (R 2 = 0.0272)

[0141] The fitting result shows that the fitting relationship between the underground dry weight of the biomass of Xanthium sibiricum and the UAV NDVI is poor, and the two have almost no correlation.

[0142] The fitting relationship between the total fresh weight of the biomass of Xanthium sibiricum and the UAV NDVI is shown in Figure 12 The fitting equation between the total fresh weight of the biomass of Xanthium sibiricum (Y) and the UAV NDVI (X) is:

[0143] Y = 103145X 2 -144810X + 53327 (R 2 = 0.4648)

[0144] The fitting result shows that the correlation coefficient R 2 of the total fresh weight of the biomass of Xanthium sibiricum and the UAV NDVI is above 0.4, and the correlation is general.

[0145] The fitting relationship between the total dry weight of the biomass of Xanthium sibiricum and the UAV NDVI is shown in Figure 13 The fitting equation between the total dry weight of the biomass of Xanthium sibiricum (Y) and the UAV NDVI (X) is:

[0146] Y = 174560X 2 -256952X + 95166 (R 2 = 0.7246)

[0147] The fitting result shows that the fitting relationship between the total dry weight of the biomass of Xanthium sibiricum and the UAV NDVI is good, and the correlation coefficient is above 0.7, and the correlation is strong.

[0148] The above analysis result shows that the aboveground fresh weight, the aboveground dry weight and the total dry weight of the biomass of Xanthium sibiricum all have very good fitting relationship with the UAV NDVI, and the correlation coefficient is above 0.7, and thus all can be used for the construction of the remote sensing inversion model of the biomass of the water-level-fluctuating zone.

[0149] 2、Dogtooth grass biomass and UAV NDVI analysis

[0150] The fitting relationship between the aboveground fresh weight of the biomass of Cynodon dactylon and the UAV NDVI is shown in Figure 14 The fitting equation between the aboveground fresh weight of the biomass of Cynodon dactylon (Y) and the UAV NDVI (X) is:

[0151] Y = 30940X 2 -36279X + 11316(R 2 = 0.5515)

[0152] The fitting results showed that the aboveground fresh weight of dogtooth grass and UAV NDVI fitting correlation coefficient R 2 was above 0.5, and the correlation was strong.

[0153] The fitting relationship between the aboveground dry weight of dogtooth grass and UAV NDVI was shown in Figure 15 The fitting equation between the aboveground dry weight of dogtooth grass biomass (Y) and UAV NDVI (X) was:

[0154] Y = 24728X 2 -28505X + 8454.4(R 2 = 0.6548)

[0155] The fitting results showed that the aboveground dry weight of dogtooth grass biomass and UAV NDVI fitting correlation coefficient R 2 was above 0.6, and the correlation was strong.

[0156] The fitting relationship between the underground fresh weight of dogtooth grass and UAV NDVI was shown in Figure 16 The fitting equation between the underground fresh weight of dogtooth grass biomass (Y) and UAV NDVI (X) was:

[0157] Y = 6464.2X 2 + 8728.8X - 2402.9(R 2 = 0.325)

[0158] The fitting results showed that the underground fresh weight of dogtooth grass biomass and UAV NDVI fitting correlation coefficient R 2 was above 0.3, and the correlation was weak.

[0159] The fitting relationship between the underground dry weight of dogtooth grass and UAV NDVI was shown in Figure 17 The fitting equation between the underground dry weight of dogtooth grass biomass (Y) and UAV NDVI (X) was:

[0160] Y = -3928.4X 2 + 5485X - 1561.2(R 2 = 0.3074)

[0161] The fitting results showed that the underground dry weight of dogtooth grass biomass and UAV NDVI fitting correlation coefficient R 2 was about 0.3, and the correlation was weak.

[0162] The fitting relationship between the total fresh weight of dogtooth grass and UAV NDVI was shown in Figure 18The fitting equation between the total fresh weight of the biomass of the dogtooth grass (Y) and the UAV NDVI (X) is:

[0163] Y=24476X 2 -27550X+8913.2(R 2 =0.4964)

[0164] The fitting result shows that the fitting correlation coefficient R between the total fresh weight of the biomass of the dogtooth grass and the UAV NDVI is about 0.5, and the correlation is relatively strong. 2

[0165] The fitting relationship between the total dry weight of the biomass of the dogtooth grass and the UAV NDVI is shown in Figure 19 The fitting equation between the total dry weight of the biomass of the dogtooth grass (Y) and the UAV NDVI (X) is:

[0166] Y=20800X 2 -23020X+6893.1(R 2 =0.6264)

[0167] The fitting result shows that the fitting correlation coefficient R between the total dry weight of the biomass of the dogtooth grass and the UAV NDVI is above 0.6, and the correlation is relatively strong. 2

[0168] The above analysis result shows that, similar to the Xanthium, the above fresh weight, the above dry weight and the total dry weight of the biomass of the dogtooth grass all have the best fitting relationship with the UAV NDVI, and the correlation coefficients are all about 0.6, and thus all can be used for the construction of the remote sensing inversion model of the biomass of the water-level-fluctuating zone.

[0169] 3. Analysis of the mixed biomass and the UAV NDVI

[0170] The fitting relationship between the above fresh weight of the mixed plant and the UAV NDVI is shown in Figure 20 The fitting equation between the above fresh weight of the biomass of the mixed plant (Y) and the UAV NDVI (X) is:

[0171] Y=13694X 2 -15468X+5525(R 2 =0.5254)

[0172] The fitting result shows that the fitting correlation coefficient R between the above fresh weight of the biomass of the mixed plant and the UAV NDVI is above 0.5, and the correlation is relatively strong. 2

[0173] The fitting relationship between the above dry weight of the mixed plant and the UAV NDVI is shown in Figure 21 The fitting equation between the above dry weight of the biomass of the mixed plant (Y) and the UAV NDVI (X) is:

[0174] ​​​Y = 977.77X 2 + 198.11X + 415.15 (R 2 = 0.3616)

[0175] The fitting results show that the fitting correlation coefficient R 2 of the aboveground biomass dry weight of mixed plants and UAV NDVI is only 0.36, and the correlation is weak.

[0176] The fitting relationship between the underground fresh weight of mixed plants and UAV NDVI is shown in Figure 22 The fitting equation of the underground biomass fresh weight (Y) of mixed plants and UAV NDVI (X) is:

[0177] Y = 2019.4X 2 - 4428.9X + 2632.6 (R 2 = 0.8049)

[0178] The fitting results show that the fitting correlation coefficient R 2 of the underground fresh weight of mixed plants and UAV NDVI is above 0.8, and the correlation is the strongest.

[0179] The fitting relationship between the underground dry weight of mixed plants and UAV NDVI is shown in Figure 23 The fitting equation of the underground biomass dry weight (Y) of mixed plants and UAV NDVI (X) is:

[0180] Y = 624.92X 2 - 1910.3X + 1281.6 (R 2 = 0.7676)

[0181] The fitting results show that the fitting correlation coefficient R 2 of the underground dry weight of mixed plants and UAV NDVI is above 0.7, and the correlation is significant.

[0182] The fitting relationship between the total fresh weight of mixed plants and UAV NDVI is shown in Figure 24 The fitting equation of the total biomass fresh weight (Y) of mixed plants and UAV NDVI (X) is:

[0183] Y = 10695X 2 - 14960X + 7887.4 (R 2 = 0.4258)

[0184] The fitting results show that the fitting correlation coefficient R 2 of the total biomass fresh weight of mixed plants and UAV NDVI is about 0.4, and the correlation is weak.

[0185] The fitting relationship between the total dry weight of mixed plants and UAV NDVI is shown in Figure 25The fitting equation of the total dry weight of mixed plant biomass (Y) and UAV NDVI (X) is:

[0186] Y = 7079.5X 2 -7925.9X + 3294 (R 2 = 0.4818)

[0187] The fitting results show that the fitting correlation coefficient R 2 of the total dry weight of mixed plant biomass and UAV NDVI is above 0.5, and the correlation is strong.

[0188] The above analysis results show that the mixed plant presents different rules from Xanthium and Cynodon. The aboveground fresh weight, underground fresh weight, and underground dry weight of the mixed plant biomass have the best fitting relationship with UAV NDVI, and the correlation coefficients all reach above 0.5. They have the strongest correlation with UAV NDVI, and thus can be used in theory for the construction of a biomass remote sensing inversion model of the water-level-fluctuating zone. However, the underground fresh weight and underground dry weight of the mixed plant biomass are not suitable for the construction of a biomass remote sensing inversion model because they have no real correlation with the formation mechanism of NDVI data.

[0189] The above correlation analysis results of Xanthium, Cynodon, and mixed plants show that the aboveground fresh weight of biomass has a strong correlation with the NDVI of the three types of vegetation. The correlation coefficient R 2 of the aboveground fresh weight of Xanthium and UAV NDVI is 0.7368, the correlation coefficient R 2 of the aboveground fresh weight of Cynodon and UAV NDVI is 0.5515, and the correlation coefficient R 2 of the aboveground fresh weight of the mixed plant and UAV NDVI is 0.5254. Therefore, the aboveground fresh weight of biomass of Xanthium, Cynodon, and the mixed plant is selected for modeling with UAV NDVI.

[0190] Three, space-based (Sentinel remote sensing) data conversion

[0191] Although the UAV data has high precision, the cost of obtaining it is huge. The space-based Sentinel remote sensing data has a spatial resolution of 10m, high resolution, and can be obtained in a large area. Therefore, it is more economical and operable to use space-based Sentinel remote sensing data to invert biomass. Therefore, it is of great practical significance to construct a conversion model between Sentinel remote sensing data and UAV NDVI and to realize dynamic observation of the biomass of the water-level-fluctuating zone based on space-based Sentinel data.

[0192] The June, July and August Sentinel data of Wushan, Zhongxian and Fuling were used as references. The corresponding month and the same place of the UAV image were selected for comparison. The same point of different land use types such as full coverage of vegetation, bare land, water area and house was selected synchronously. Since the spatial resolution of Sentinel data is 10 m and the spatial resolution of UAV is 0.1 m, the average value of NDVI in 100*100 pixel range was calculated for the UAV image to match the Sentinel data.

[0193] Ten to fifteen points were selected for each land use type, and a total of 168 points were selected. Among them, 50 points were selected in Fuling, 54 points were selected in Zhongxian, and 64 points were selected in Wushan.

[0194] The following are the fitting processes of different months in three places.

[0195] The fitting relationship between Fuling August Sentinel NDVI (X) and UAV NDVI (Y) is shown in Figure 26 . The fitting result of Fuling August Sentinel NDVI (X) and UAV NDVI (Y) is:

[0196] Y = 1.5324X - 0.3392 (R 2 = 0.9483)

[0197] The fitting result shows that the fitting relationship between Fuling August Sentinel NDVI and its corresponding UAV NDVI is very good, the correlation coefficient R 2 = 0.9483, and the correlation is extremely significant.

[0198] The fitting relationship between Zhongxian July Sentinel NDVI (X) and UAV NDVI (Y) is shown in Figure 27 . The fitting result of Zhongxian July Sentinel NDVI (X) and UAV NDVI (Y) is:

[0199] Y = 1.3248X - 0.3 (R 2 = 0.954)

[0200] The fitting result shows that the fitting relationship between Zhongxian July Sentinel NDVI and its corresponding UAV NDVI is very good, the correlation coefficient R 2 = 0.954, and the correlation is extremely significant.

[0201] The fitting relationship between Wushan June Sentinel NDVI and UAV NDVI is shown in Figure 28 . The fitting result of Wushan June Sentinel NDVI (X) and UAV NDVI (Y) is:

[0202] Y = 1.3256X - 0.2947 (R 2 = 0.9675)

[0203] The fitting results show that the NDVI of the June Sentinel in Wushan and the corresponding UAV NDVI have a very good fitting relationship, with a correlation coefficient R 2 = 0.9675, and a very significant correlation.

[0204] The fitting relationship of the August Sentinel NDVI and the UAV NDVI data in Fuling, the July Sentinel NDVI and the UAV NDVI data in Zhongxian, and the June Sentinel NDVI and the UAV NDVI data in Wushan is shown in Figure 29 The fitting results of the August Sentinel NDVI (X) and the UAV NDVI (Y) in Fuling, the July Sentinel NDVI (X) and the UAV NDVI (Y) in Zhongxian, and the June Sentinel NDVI (X) and the UAV NDVI (Y) in Wushan are as follows:

[0205] Y = 1.3695X - 0.3007 (R 2 = 0.954)

[0206] The comprehensive fitting results show that the NDVI of the August Sentinel in Fuling, the July Sentinel in Zhongxian, and the June Sentinel in Wushan and the corresponding UAV NDVI have a very good fitting relationship, with a correlation coefficient R 2 = 0.954, and a very significant correlation. Therefore, the space-based Sentinel NDVI data can be converted into air-based UAV NDVI remote sensing data using this comprehensive fitting formula to improve the accuracy of biomass inversion.

[0207] IV. Prospects

[0208] 1. Conducting vegetation classification in the water-storage belt

[0209] During the research process, if all the sample plots are considered as one type, it will be found that the correlation between biomass and UAV NDVI is too poor to build an inversion model. Therefore, after dividing the vegetation types in the water-storage belt into three types, i.e., Xanthium, Cynodon, and mixed types, respectively, the inversion model can achieve better results. Therefore, conducting hyperspectral research on the vegetation classification in the water-storage belt can greatly improve the accuracy of biomass inversion in the water-storage belt.

[0210] 2. Improving the performance of monitoring equipment

[0211] The multi-spectral UAV used in this study does not have the RTK precise positioning function due to the limitations of the equipment itself. As shown in Figure 30 , the square frame is the positioning of the UAV, and the dark block in the image is the actual positioning of the sample plot. It can be seen that there is a deviation between the two, and there is a positioning error of about 1 meter in space when matching the UAV image data with the sample plot measured points. In this study, spatial correction was performed by using the difference in ground cover before and after the sample plot was harvested to ensure the spatial matching accuracy, but the efficiency is not high and it is easy to make mistakes. Therefore, a better UAV device, such as a hyperspectral-radar integrated machine, can not only accurately position, but also obtain vegetation height, terrain, and other model parameters, which can have a more ideal simulation effect on the sample plot measured biomass remote sensing modeling.

[0212] The above-described embodiments are merely preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Any equivalent substitutions or transformations made by those skilled in the art based on the present application are within the protection scope of the present application. The protection scope of the present application is subject to the claims.

Claims

1. A method for integrated air-ground monitoring of typical vegetation biomass in the drawdown zone of the Three Gorges Reservoir area, characterized in that, It includes the following steps: Following the principles of uniform distribution, consideration of topography and landforms, good flight conditions, and diverse land cover types, and taking into account the topography of the drawdown zone, the vegetation types within the drawdown zone, the external conditions of the drawdown zone, and location factors, the Three Gorges Reservoir drawdown zone area, which is representative, universal, operable, and typical, was selected as the study area. Using a combination of satellite remote sensing, UAV aerial photography, and ground measurement, with the Normalized Difference Vegetation Index (NDVI) as the core, the study area was subjected to sky-air-ground vegetation index extraction and observation in four time periods: May, June, July, and August. Finally, a low-to-medium-to-high resolution typical vegetation biomass inversion model of the drawdown zone was established based on satellite NDVI, UAV NDVI, and ground measurement data, and the spatiotemporal dynamics of biomass changes during the vegetation growing season from May to August in the drawdown zone were analyzed. The satellite remote sensing uses the European Sentinel-2 satellite; the drone aerial photography uses the DJI Phantom 4 Multispectral Edition drone; the ground measurement uses ground quadrat or transect surveys, which are conducted when the vegetation is fully grown. The survey content includes the name of the ground vegetation, land use type, height, biomass, and distribution location. At the same time, a field crop NDVI measuring instrument is used to measure the ground NDVI. In order to improve the accuracy of typical vegetation biomass inversion in the drawdown zone during ground quadrat or transect surveys, the ground measured data were divided into three groups: Bermuda grass, Xanthium sibiricum, and mixed, and analyzed separately. The mixed group refers to the vegetation types inside the drawdown zone other than Bermuda grass and Xanthium sibiricum. The biomass is the aboveground fresh weight; the study area consists of five regions: Xuetangwan in Wushan, Dalangba in Kaizhou, Shibangou in Yunyang, Hongyanzi in Zhongxian, and Muhe Village in Fuling, covering two tributary points and three points along the main stream of the Yangtze River in the Three Gorges Reservoir area; The typical vegetation biomass inversion model of the drawdown zone includes an inversion model between ground-measured biomass data and UAV NDVI data, and an inversion model between Sentinel NDVI and UAV NDVI. In the inversion model between ground-measured biomass data and UAV NDVI data, the fitting equation for the aboveground fresh weight (Y) of Xanthium sibiricum biomass and UAV NDVI (X) is as follows: Y=224062X 2 -331611X+124488(R 2 =0.7368); The fitting equation for the aboveground fresh weight (Y) of bermudagrass biomass versus UAV NDVI (X) is as follows: Y=30940X 2 -36279X+11316(R) 2 =0.5515) ; The fitting equation for the aboveground biomass fresh weight (Y) of mixed plants and the NDVI (X) of the UAV is: Y=13694X 2 -15468X+5525(R 2 =0.5254); In the inversion model between Sentinel NDVI and UAV NDVI, the fitting equations for Sentinel NDVI (X) and UAV NDVI (Y) are as follows: Y=1.3695X-0.3007 (R 2 =0.954)。 2. The method for integrated air-ground monitoring of typical vegetation biomass in the drawdown zone of the Three Gorges Reservoir area according to claim 1, characterized in that, The drawdown zone topography is one or more of the following: steep slope, flat plain, gentle slope, and reservoir tail type; the vegetation type inside the drawdown zone is one or more of the following: cocklebur, bermudagrass, wolfberry, flat-spike bullwhip, nutgrass, short-leaved water centipede, and ragweed; the external conditions of the drawdown zone are land use types, including one or more of the following: cultivated land, forest land, shrubland, orchard, garden land, grassland, bare land, water area, and settlement.

3. The method for integrated air-ground monitoring of typical vegetation biomass in the drawdown zone of the Three Gorges Reservoir area according to claim 1, characterized in that, During ground monitoring, to ensure data quality, there should be no fewer than 5 ground NDVI monitoring points and quadrat points and no fewer than 1 transect in each study area, ensuring that there are no fewer than 10 points per 1 km² in each study area, and that 5 sets of values ​​are measured at each point within a 1 M × 1 M area and the average value is taken.

4. The method for integrated air-ground monitoring of typical vegetation biomass in the drawdown zone of the Three Gorges Reservoir area according to claim 1, characterized in that, The method for constructing the inversion model between the biomass data obtained from ground measurements and the UAV NDVI data is as follows: UAV NDVI data at the corresponding time and space are obtained based on the ground measurement points, and then correlation linear analysis is performed with the aboveground fresh weight, aboveground dry weight, underground fresh weight, underground dry weight, total fresh weight, and total dry weight of Xanthium sibiricum, Cynodon dactylon, and mixed sample plots to obtain the relationship between biomass and NDVI for different vegetation types. Finally, considering the biological significance and the rationality of the correlation, the inversion model between the biomass data obtained from ground measurements and the UAV NDVI data is constructed.

5. The integrated air-ground monitoring method for typical vegetation biomass in the drawdown zone of the Three Gorges Reservoir area according to claim 1, characterized in that, The method for constructing the inversion model between sentinel NDVI and UAV NDVI is as follows: using sentinel data from different months in the study area as a reference, UAV images of the corresponding months and the same locations are selected for comparison, and corresponding points under different land use types such as full vegetation cover, bare land, water area, and buildings are selected simultaneously; since the spatial resolution of sentinel data is 10m and the spatial resolution of UAV is 0.1m, the average NDVI of UAV images within a 100×100 pixel range is calculated to match the sentinel data.

Images