CN115526098B - Remote sensing calculation method for leaf area index of surface vegetation in mining area and electronic equipment - Google Patents
Remote sensing calculation method for leaf area index of surface vegetation in mining area and electronic equipment Download PDFInfo
- Publication number
- CN115526098B CN115526098B CN202211120240.6A CN202211120240A CN115526098B CN 115526098 B CN115526098 B CN 115526098B CN 202211120240 A CN202211120240 A CN 202211120240A CN 115526098 B CN115526098 B CN 115526098B
- Authority
- CN
- China
- Prior art keywords
- vegetation
- leaf
- remote sensing
- model
- area index
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000005065 mining Methods 0.000 title claims abstract description 52
- 238000004364 calculation method Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 69
- 238000013528 artificial neural network Methods 0.000 claims abstract description 54
- 238000012549 training Methods 0.000 claims abstract description 44
- 230000008878 coupling Effects 0.000 claims abstract description 24
- 238000010168 coupling process Methods 0.000 claims abstract description 24
- 238000005859 coupling reaction Methods 0.000 claims abstract description 24
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 23
- 238000005259 measurement Methods 0.000 claims abstract description 19
- 238000001228 spectrum Methods 0.000 claims abstract description 18
- 238000012360 testing method Methods 0.000 claims abstract description 15
- 238000004088 simulation Methods 0.000 claims description 40
- 239000002689 soil Substances 0.000 claims description 39
- 230000008569 process Effects 0.000 claims description 34
- 238000002310 reflectometry Methods 0.000 claims description 30
- 230000006870 function Effects 0.000 claims description 29
- 229930002875 chlorophyll Natural products 0.000 claims description 20
- 235000019804 chlorophyll Nutrition 0.000 claims description 20
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 claims description 20
- 230000003595 spectral effect Effects 0.000 claims description 20
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 15
- 229930002877 anthocyanin Natural products 0.000 claims description 13
- 235000010208 anthocyanin Nutrition 0.000 claims description 13
- 239000004410 anthocyanin Substances 0.000 claims description 13
- 150000004636 anthocyanins Chemical class 0.000 claims description 13
- 239000001058 brown pigment Substances 0.000 claims description 13
- 235000021466 carotenoid Nutrition 0.000 claims description 12
- 150000001747 carotenoids Chemical class 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 12
- 238000011160 research Methods 0.000 claims description 10
- 238000005316 response function Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 9
- 238000009826 distribution Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 7
- 241000257303 Hymenoptera Species 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 abstract description 4
- 239000000523 sample Substances 0.000 description 9
- 230000015572 biosynthetic process Effects 0.000 description 7
- 238000003786 synthesis reaction Methods 0.000 description 7
- 238000010801 machine learning Methods 0.000 description 6
- 238000001035 drying Methods 0.000 description 5
- 238000007637 random forest analysis Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 239000003016 pheromone Substances 0.000 description 4
- 238000005303 weighing Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 238000004321 preservation Methods 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000009827 uniform distribution Methods 0.000 description 3
- 239000002028 Biomass Substances 0.000 description 2
- 102000004190 Enzymes Human genes 0.000 description 2
- 108090000790 Enzymes Proteins 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000009849 deactivation Effects 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004836 empirical method Methods 0.000 description 2
- 238000004806 packaging method and process Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- LFQSCWFLJHTTHZ-UHFFFAOYSA-N Ethanol Chemical compound CCO LFQSCWFLJHTTHZ-UHFFFAOYSA-N 0.000 description 1
- UPYKUZBSLRQECL-UKMVMLAPSA-N Lycopene Natural products CC(=C/C=C/C=C(C)/C=C/C=C(C)/C=C/C1C(=C)CCCC1(C)C)C=CC=C(/C)C=CC2C(=C)CCCC2(C)C UPYKUZBSLRQECL-UKMVMLAPSA-N 0.000 description 1
- 238000012952 Resampling Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 150000001746 carotenes Chemical class 0.000 description 1
- 235000005473 carotenes Nutrition 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000029553 photosynthesis Effects 0.000 description 1
- 238000010672 photosynthesis Methods 0.000 description 1
- 230000009024 positive feedback mechanism Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000000179 transient infrared spectroscopy Methods 0.000 description 1
- 230000005068 transpiration Effects 0.000 description 1
- NCYCYZXNIZJOKI-UHFFFAOYSA-N vitamin A aldehyde Natural products O=CC=C(C)C=CC=C(C)C=CC1=C(C)CCCC1(C)C NCYCYZXNIZJOKI-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Pure & Applied Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Business, Economics & Management (AREA)
- Marine Sciences & Fisheries (AREA)
- Primary Health Care (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- Agronomy & Crop Science (AREA)
- Animal Husbandry (AREA)
- Geometry (AREA)
- Mining & Mineral Resources (AREA)
- Computational Linguistics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
Abstract
The invention discloses a remote sensing calculation method for leaf area index of surface vegetation in a mining area and an electronic device, wherein the method comprises the following steps: s1, establishing a vegetation parameter inversion model by using the constructed PROSAIL coupling model to couple remote sensing data of a satellite sensor and combining ground actual measurement spectrum and parameter data; s2, building an artificial depth neural network containing a vegetation parameter inversion model, and applying an ant colony algorithm to the artificial depth neural network for training and testing by taking a mean square error as an adaptability value of an ant and taking a shortest path determined by an ant colony as an optimal initial weight and an optimal bias parameter; and S3, inputting the mining area earth surface vegetation remote sensing data through a vegetation parameter inversion model, and then outputting a leaf area index. The invention carries out remote sensing monitoring on the leaf area index with long time sequence and high space-time resolution in the mining area, improves the inversion precision of the leaf area index, provides basis for informatization of vegetation condition supervision in the mining area, and provides decision technical support for ecological management and ecological restoration in the mining area.
Description
Technical Field
The invention relates to the field of vegetation leaf area index inversion, in particular to a remote sensing calculation method for vegetation leaf area indexes of ground surfaces of mining areas and electronic equipment.
Background
Coal mining can cause a series of influences on the ecological environment of mining areas, and form a land vegetation pattern with mining activity characteristics and a change process. The leaf area index represents the density degree and the canopy structure characteristics of the leaves, embodies the capability of biological physical processes such as vegetation photosynthesis, respiration and transpiration, and is a key parameter for describing the exchange of substances and energy between soil, vegetation and atmosphere. The method for estimating the leaf area index by using the optical remote sensing data mainly comprises two major categories, wherein the first category is an empirical method, namely, the empirical statistical relationship between the reflectivity observed by remote sensing or the vegetation index calculated by using the reflectivity and the leaf area index measured on the ground is directly established, and then the empirical statistical relationship is popularized to a region to complete regional leaf area index mapping. The method has strong applicability and high precision in a small area, but the ground observation needs to consume a large amount of manpower and material resources, is greatly influenced by regional background conditions and vegetation types, cannot be directly popularized in a large range as a general method, and the mining area leaf area index cannot be obtained through an empirical method because the mining area leaf area index cannot be calculated through long-time-sequence high-space-time frequency remote sensing without prior data. The second type is a physical process, which is generally divided into two steps: the method comprises the first step of simulating the radiation transmission process of photons in the canopy based on the photon transmission theory of the vegetation canopy, establishing a forward simulation model of leaf area index and other leaf, canopy and background biophysical parameters and canopy spectral reflectivity, and the second step of estimating spectral contribution parameters such as the leaf area index and the like by taking the earth surface reflectivity obtained by remote sensing and other earth surface known information as input on the basis of the established forward simulation model and performing reverse calculation.
Because of the relatively numerous unknown canopy parameters, the information provided by remote sensing observation is limited, and therefore, a unique analytical solution cannot be obtained by inversion generally. Aiming at the problem, in the second-step inversion process, an inversion result closest to a theoretical true value is obtained more conveniently and efficiently by means of an optimization algorithm, a lookup table method and some machine learning algorithms. The optimization algorithm is flexible to use, can synchronously start operation from a plurality of initial values and even synchronously optimize on a plurality of pixels, has high inversion precision, but is easy to fall into a local minimum value in the traditional optimization algorithm, so that a globally optimal result is not obtained, and secondly, a large amount of computer operation resources are consumed in the optimization process, and the consumed time is large. The lookup table method enables the LAI inversion process to be more efficient, an initial value is not required to be given, the globally optimal solution can be obtained, the dimensionality of the lookup table needs to be large enough to meet the requirement of inversion accuracy, the sampling interval of variables needs to be small enough, the inversion speed can be reduced, and if the parameters are not simplified properly, the applicability of the lookup table method inversion can be restricted. The most commonly used Artificial Neural Network (ANN) in machine learning algorithms, the performance of ANN depends on the design of its structure, while too many or too few layers or neuron settings may significantly reduce its accuracy, and the creation of network results and parameter adjustment require large computational resources and time consumption when applied to large data sets.
Ensemble learning is a method for building a more powerful model by combining a plurality of machine learning models, and in ensemble learning, a most commonly used LAI remote sensing inversion method is used in Random Forest (RF) regression based on decision tree (decision tree). Compared with an Artificial Neural Network (ANN), random Forest (RF) generally can obtain better results without scaling data and repeatedly adjusting parameters, so that the stability of the LAI inversion result is greatly improved. However, the process of constructing a random forest is still time-consuming for large data, and the estimation accuracy is difficult to continuously improve through continuous parameter adjustment. There is also a kernel method for machine learning, the most commonly used kernel methods are Support Vector Machine (SVM) and Gaussian Process Regression (GPR), and since there are a lot of hyper-parameters in the model to be adjusted, the SVM method also consumes a lot of time during training. The GPR is slightly improved in inversion accuracy compared with other machine learning methods, and the model training speed is higher. However, the GPR method is currently only applied to a small training data set for regional LAI-oriented inversion and is not applied to business.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a remote sensing calculation method for the leaf area index of the vegetation on the ground surface of the mining area and electronic equipment, which are used for carrying out remote sensing monitoring on the leaf area index with long time sequence and high space-time resolution, improve the inversion precision of the leaf area index, provide a basis for informatization of supervision of the vegetation condition of the mining area and provide decision technical support for ecological management and ecological restoration of the mining area.
The purpose of the invention is realized by the following technical scheme:
a remote sensing calculation method for leaf area indexes of surface vegetation in a mining area comprises the following steps:
s1, coupling a PROSAIL coupling model through a PROSPECT model and an SAILH model, and establishing a vegetation parameter inversion model based on a full-connection network model by adopting the PROSAIL coupling model to couple remote sensing data of a satellite sensor and combining ground actual measurement spectrum and parameter data; input parameters of the PROSAIL coupling model comprise blade dimension, canopy dimension, background soil and observation geometry;
s2, a vegetation parameter inversion model produces a long-time sequence vegetation parameter product on a Google Earth Engine platform, an artificial deep neural network comprising the vegetation parameter inversion model is built, the artificial deep neural network adopts an ant colony algorithm, takes a mean square error as an adaptability value of ants, and takes a shortest path determined by ant colonies as an optimal initial weight and an optimal bias parameter to be given to the artificial deep neural network for training and testing;
and S3, inputting the mining area earth surface vegetation remote sensing data through a vegetation parameter inversion model, and then outputting a leaf area index.
In order to better realize the remote sensing calculation method of the area index of the earth surface vegetation in the mining area, the input data of the PROSAIL coupling model comprises a leaf scale, a canopy scale, background soil and observation geometry, the leaf scale comprises chlorophyll content, dry matter content, leaf structure parameters, carotenoid content, brown pigment content, anthocyanin content and leaf equivalent water thickness, the chlorophyll content, the dry matter content, the leaf equivalent water thickness, the carotenoid content, the brown pigment content and the anthocyanin content are measured by sampling in the mining area range, the initial parameters of the carotenoid content, the brown pigment content and the anthocyanin content are set to be zero, and the range of the leaf structure parameter value is set to be between 1 and 2; the canopy scale comprises a leaf area index and an average leaf inclination angle, and the leaf area index and the average leaf inclination angle are obtained by field measurement; the background soil comprises the proportion of dry soil and soil spectrum data, the proportion of the dry soil is measured by sampling in a mining area, and the soil spectrum data is obtained by field measurement; the observation geometry comprises a solar zenith angle, an observation zenith angle and a relative azimuth angle, and the observation geometry is derived from remote sensing data of a satellite sensor; the chlorophyll content, the dry matter content, the leaf equivalent water thickness, the carotenoid content, the brown pigment content, the anthocyanin content, the proportion of dry soil and the soil spectrum data are measured for multiple times to determine the numerical distribution, the average value, the maximum value and the minimum value of the parameters.
Preferably, the remote sensing data of the satellite sensor is from a Landsat series satellite sensor or a Sentinel-2A satellite sensor; the remote sensing data of the satellite sensor comprises six wave bands B1, B2, B3, B4, B5 and B7, the directional reflectivity of the canopy is obtained according to the PROSAIL coupling model, and the directional reflectivity is input into the vegetation parameter inversion model in combination with the spectral response function of the remote sensing data of the satellite sensor for training.
Preferably, the remote sensing data of the satellite sensor is derived from a Landsat series satellite sensor, the Landsat series satellite sensor includes Landsat5, landsat7 and Landsat8, and the remote sensing data of the satellite sensor is subjected to registration processing with the Landsat8 as a reference according to the following method:
selecting a small research area, extracting pixel values of the research area to points, respectively constructing regression linear models among corresponding pixels of different images for Landsat5, landsat7 and Landsat8 by using the least square principle, and realizing registration with Landsat8 as a reference based on the regression linear models.
Preferably, the vegetation parameter inversion model comprises a forward simulation model and a reverse simulation model, the forward simulation model realizes a forward simulation process from inputting the leaf area index to outputting the directional reflectivity of the canopy, the reverse simulation model takes the directional reflectivity of the canopy as input and takes the prior leaf area index as output, and the vegetation parameter inversion model continuously adjusts and optimizes the model parameters according to the difference between the minimum estimated leaf area index and the prior leaf area index through the training process of the forward simulation model and the reverse simulation model.
The optimized artificial deep neural network calculates a loss function according to the leaf area index obtained by simulation and the leaf area index obtained by actual measurement, and optimizes the weight and the bias parameter by using the loss value and the optimization function.
The preferred artificial deep neural network training process of the invention is as follows:
s21, carrying out standardization processing on simulation data obtained by the vegetation parameter inversion model and converting the simulation data into dimensionless data; the normalization process uses the following formula:
wherein x represents the normalized value, u represents the mean, and σ represents the standard deviation;
s22, randomly dividing the data after the standardization treatment into a training set and a testing set according to the ratio of 7:3, and training the artificial deep neural network by using the training set to obtain a topological structure of the artificial deep neural network; the loss function of the artificial deep neural network is calculated by adding the following two loss functions:
S=a*L 1 +b*L 2 wherein L is 1 Representing a network intermediate loss function which is the error between a process value and a real value; l is 2 Representing the final loss function of the network, wherein the final loss function is the error between the output result and the true value, and a and b are weights respectively;
network intermediate loss function L 1 Net final loss function L 2 All adopt the following root mean square error loss functionNumber:
and S23, obtaining the trained artificial deep neural network through training, and testing the trained artificial deep neural network by using a test set.
Preferably, in the method S22, the artificial deep neural network further adopts an adam optimization algorithm to obtain higher computational efficiency and lower memory requirement.
Preferably, the artificial deep neural network comprises an input layer, a hidden layer and an output layer, wherein the input layer is used for inputting the vegetation parameter inversion training data set, the hidden layer is used for ant colony algorithm processing, and the output layer is used for outputting the leaf area index.
An electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the steps of the method for calculating remote sensing of mine area vegetation leaf area index of the present invention.
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) The invention carries out remote sensing monitoring on the leaf area index with long time sequence and high space-time resolution in the mining area, improves the inversion precision of the leaf area index, provides basis for informatization of vegetation condition supervision in the mining area, and provides decision technical support for ecological management and ecological restoration in the mining area.
(32) The method is based on auxiliary data such as a data set for field data acquisition and PROSAIL coupling model training, landsat series satellites and the like, an artificial deep learning neural network structure optimized by an ant colony algorithm is provided for the first time, a technical method for leaf area index inversion is constructed, and leaf area index data with long time sequence and high space-time frequency in a mine area are obtained; the invention adopts a positive feedback mechanism, so that the weight and the bias approach to an optimal solution, and the disadvantages of large consumption of computer operation resources, large calculation dimension, small applicable area of an optimization algorithm, a lookup table method and a machine learning method are avoided.
(3) According to the method, based on field acquisition data, a data set trained by a PROSAIL coupling model, landsat remote sensing image data and other auxiliary data, the LAI data with the resolution of 30 meters in the mining area are obtained by combining the scene of the mining area with the basic information of the mining area; by means of sensor correction, the LAI calculated by different satellites in 1990-2020 is unified to a Landsat-8 standard, a leaf area index product with consistent long-sequence high space-time resolution and adaptation to a mining area scene is obtained, support is provided for informatization of vegetation condition supervision of the mining area, and decision support is provided for ecological management and ecological restoration of the mining area.
(4) The method obtains the value of the input parameter of the PROSAIL coupling model through field measurement, and obtains the value range, the average value, the maximum value and the minimum value of the input parameter of the PROSAIL coupling model suitable for the mining area environment through calculation, thereby effectively improving the inversion precision of the leaf area index.
(5) According to the method, vegetation parameter inversion model forward modeling is utilized to obtain canopy reflectivity data of vegetation of 400-2500nm, and then a spectral response function of a Landsat-5-7-8 satellite is combined for inversion to obtain a mining area leaf area index training data set, so that data support is provided for remote sensing inversion of the mining area long-time sequence high space-time resolution leaf area index.
Drawings
FIG. 1 is a schematic flow chart of a remote sensing calculation method of leaf area index according to an embodiment of the present invention;
FIG. 2 is a structural diagram of an artificial deep neural network model optimized based on an ant colony algorithm in the embodiment;
FIG. 3 is a schematic diagram of forward simulation and reverse simulation in the embodiment;
FIG. 4 is a schematic diagram of the inversion of the leaf area index in an embodiment;
fig. 5 is a diagram illustrating the LAI effect of the mining area according to an example.
Detailed Description
The present invention will be described in further detail with reference to the following examples:
examples
As shown in fig. 1, a remote sensing calculation method for leaf area index of surface vegetation in a mining area comprises the following steps:
s1, coupling a PROSAIL coupling model through a PROSPECT model and an SAILH model, and establishing a vegetation parameter inversion model based on a full-connection network model by adopting the PROSAIL coupling model to couple remote sensing data of a satellite sensor and combining ground actual measurement spectrum and parameter data; input parameters of the PROSAIL coupling model include blade dimensions, canopy dimensions, background soil, and observation geometry. The input data of the PROSAIL coupling model comprises leaf scale, canopy scale, background soil and observation geometry, wherein the leaf scale comprises chlorophyll content, dry matter content, leaf structure parameters, carotenoid content, brown pigment content, anthocyanin content and leaf equivalent water thickness, and the chlorophyll content, dry matter content, leaf equivalent water thickness, carotenoid content, brown pigment content and anthocyanin content are measured by sampling in a mining area. For example: the chlorophyll content is collected on the spot, the chlorophyll content of the vegetation is measured by adopting a portable novel chlorophyll instrument SPAD-502, the relative content value of the chlorophyll content of the vegetation canopy is obtained by using the portable novel chlorophyll instrument SPAD, and the size of the measured value represents the chlorophyll content of the leaves in unit area. In order to obtain the chlorophyll content, the SPAD value of the vegetation is obtained at three different positions of the vegetation by using a portable novel chlorophyll meter SPAD, each sample point is measured for 3 times, and the average value is the final SPAD value of the leaf. The chlorophyll content Cab of the vegetation leaves is converted by utilizing the measured SPAD value of the leaves, and the formula is as follows: cab =0.11SPAD 1.5929 . The initial parameters of the carotenoid content, the brown pigment content and the anthocyanin content are set to be zero, and the range of the leaf structure parameter value is set to be between 1 and 2. In some embodiments, the dry matter content is obtained as follows: collecting biomass of sample point, collecting vegetation leaves at upper, middle and lower parts, packaging into plastic bags, adding ice in thermal insulation box for low temperature preservation, and rapidly returning to realA laboratory; opening the vegetation blades in the plastic package bag in a laboratory, flattening, wiping the blades clean, then measuring the length and the width of a single blade by using a tape measure, measuring each blade for three times, taking an average value as a final length and width value, and recording according to corresponding numbers. Numbering the vegetation leaves of each sample point, placing the vegetation leaves in different areas, carrying out one-hour enzyme deactivation treatment at 105 ℃, then drying the vegetation leaves at 80 ℃, stopping drying when the weighing error is less than 0.005g through multiple measurements, and weighing. The dry matter content of the vegetation leaves is the ratio of the dry matter mass of the vegetation canopy leaves subjected to indoor drying to the leaf area of the vegetation leaves, and Cm = W/S, wherein Cm is the dry matter content of the leaves, W is the dry weight of the leaves, and S is the leaf area of the leaves.
The canopy scale comprises a leaf area index, an average leaf inclination angle, and the leaf area index and the average leaf inclination angle are obtained by field measurement; the background soil comprises the proportion of dry soil and soil spectrum data, the proportion of the dry soil is measured by sampling in a mining area, and the soil spectrum data is obtained by field measurement; in the embodiment, soil spectral data can be collected on site, a spectral radiometer SVC is used for measuring the spectral reflectivity of the soil in a mining area, clear, cloudless and windless dates are selected, two points from ten am to two points in the afternoon are selected for measuring time, the solar altitude angle is greater than 60 degrees, the spectral field angle is set to be 25 degrees, a portable tripod is used for mounting the spectral radiometer on the tripod, a sensor is arranged right opposite to the observed soil, the sensor probe is downward, the height from the probe to the soil is about 0.5 meter, and in order to reduce the interference of environmental factors, a reference white board needs to be corrected before each measurement. And measuring the spectral reflectivity of 3 pieces of soil at each sample point, and averaging to obtain the final reflectivity of the soil. In the embodiment, soil spectral data can be collected on site, a vegetation leaf area index LAI is obtained by adopting an LAI-2200 instrument, the measurement is carried out on clear, cloudless and windless dates, the measured time is two points from ten am to afternoon, the solar altitude angle is more than 60 degrees, a file folder for storing observation data is built before measurement, the canopy is measured once, the root is measured four times, each sample point is measured five times in total, and the average value of the five results is taken as a leaf area index measured value. The observation geometry comprises a solar zenith angle, an observation zenith angle and a relative azimuth angle, and the observation geometry is derived from remote sensing data of a satellite sensor. The chlorophyll content, the dry matter content, the leaf equivalent water thickness, the carotenoid content, the brown pigment content, the anthocyanin content, the proportion of dry soil and the soil spectrum data are measured for multiple times to determine the numerical distribution, the average value, the maximum value and the minimum value of the parameters; in order to cover the different vegetation types (such as shrubs and grasslands) in the mining area as widely as possible, the value distribution range of most parameters is determined before the PROSAIL coupling model processing.
The leaf equivalent water thickness is the difference between the fresh weight mass of the vegetation and the dried dry matter mass in unit area, and then the difference is compared with the leaf area of the vegetation; in some embodiments, the blade equivalent water thickness may be obtained by: collecting biomass of a sample point, mainly comprising an upper part, a middle part and a lower part, collecting vegetation leaves, filling the vegetation leaves into a plastic packaging bag, adding ice blocks into a heat preservation box for low-temperature preservation, and quickly taking the vegetation leaves back to a laboratory; opening the vegetation blades in the plastic package bag in a laboratory, flattening, wiping the blades clean, then measuring the length and the width of a single blade by using a tape measure, measuring each blade for three times, taking an average value as a final length and width value, and recording according to corresponding numbers. Numbering the vegetation leaves of each sample point, putting the vegetation leaves in different areas, carrying out one-hour enzyme deactivation treatment at 105 ℃, then drying the vegetation leaves at 80 ℃, stopping drying when the weighing error is less than 0.05g through multiple measurements, and weighing. The calculation formula of the equivalent water thickness EWT of the blade is as follows: EWT = (W) f -W d ) (S), wherein EWT is blade equivalent water thickness, W f Is the fresh weight of the leaf, W d Is the dry weight of the leaf and S is the area of the leaf.
S2, the vegetation parameter inversion model produces a long-time sequence vegetation parameter product on a Google Earth Engine platform, an artificial deep neural network containing the vegetation parameter inversion model is built, the artificial deep neural network adopts an ant colony algorithm, takes the mean square error as the adaptability value of ants, and takes the shortest path determined by ant colonies as the optimal initial weight and bias parameters to be given to the artificial deep neural network for training and testing. The invention extracts the weight and bias elements in the network to form the path coordinates of the ant population; since the shorter the path of an ant to a food source, the higher the pheromone content on the path, the mean square error is taken as the fitness value of the ant. And finally, the shortest path determined by the ant population is used as the optimal initial weight and the optimal bias parameter, and then is given to the artificial deep neural network for testing and training. In some embodiments, the ant colony algorithm optimization artificial deep neural network parameter method is as follows: s1, reading data, and initializing the structure of the artificial deep neural network and parameters of an ant colony algorithm. S2, calculating the dimension of a solution space, and initializing ant positions and the highest pheromone. And S3, calculating the content of the pheromone according to the positions of the ants. And S4, calculating the highest pheromone and updating the optimal individual position. And S5, transferring and updating the ant positions according to the probability. And S6, executing a loop body from S3 to S5 to reach a termination algebra. And S7, taking out the optimized optimal ant position coordinate, and giving the optimal ant position coordinate to the artificial neural network to obtain the optimal initial weight and bias. And S8, training and testing the optimized artificial neural network. Preferably, the artificial deep neural network comprises an input layer, a hidden layer and an output layer, wherein the input layer is used for inputting the vegetation parameter inversion training data set, the hidden layer is used for ant colony algorithm processing, and the output layer is used for outputting the leaf area index.
In some embodiments, as shown in fig. 3, the vegetation parameter inversion model includes a forward simulation model and a reverse simulation model, the forward simulation model implements a forward simulation process from inputting the leaf area index to outputting the directional reflectivity of the canopy, the reverse simulation model uses the directional reflectivity of the canopy as input and uses the prior leaf area index as output, and the vegetation parameter inversion model continuously adjusts and optimizes the model parameters according to the difference between the minimum estimated leaf area index and the prior leaf area index through the training processes of the forward simulation model and the reverse simulation model. The vegetation parameter inversion model establishes a forward simulation process from the leaf area index LAI and other vegetation parameters to the reflectivity, and the leaf area index LAI remote sensing inversion reversely infers a corresponding leaf area index LAI result, namely the reverse process of the forward simulation process, based on the reflectivity and other angles and other information observed by the sensor. The basic processes of forward simulation and reverse inversion of the LAI (i.e. leaf area index) remote sensing estimation are shown in fig. 3, where the left-hand part is the forward simulation process and the right-hand part is the LAI inversion process. The basic process of the LAI inversion of the model training method is shown in fig. 4, and in fig. 4, the left side is a simulation process from the leaf area index LAI to the canopy reflectivity from bottom to top, that is, the construction of a training data set is completed; and then, training an inversion model by taking the simulated canopy reflectivity as input and the prior LAI as output, and continuously adjusting model parameters in the training process by minimizing the difference between the estimated LAI and the prior LAI to finally obtain the optimal model parameters and structure. And during inversion, inputting the observed reflectivity into a trained model, wherein the output LAI of the model is an inversion result.
In some embodiments, the artificial deep neural network calculates a loss function according to the leaf area index obtained by simulation and the leaf area index obtained by actual measurement, and optimizes the weight and the bias parameter by using the loss value and the optimization function. In some embodiments, the artificial deep neural network training process is as follows:
s21, carrying out standardization processing on simulation data obtained by the vegetation parameter inversion model and converting the simulation data into dimensionless data; the normalization process uses the following formula:
wherein x represents the normalized value, u represents the mean, and σ represents the standard deviation;
s22, the data after the standardization treatment are processed according to the following steps of 7:3, randomly dividing the artificial deep neural network into a training set and a testing set, and training the artificial deep neural network by using the training set to obtain a topological structure of the artificial deep neural network; for example: as shown in FIG. 2, the input layer of the present invention includes 6 neurons corresponding to 6 bands of the sensor, respectively, and the output layer has only one neuron corresponding to the LAI. Inputting 6 wave band variables to carry out network calculation, finally determining the structure of the artificial deep neural network optimized by the ant colony algorithm to be 12-12-12-36-12-12-12-6 through a plurality of experiments by a trial-and-error method, wherein the structure of the artificial deep neural network is eight layers in total, and the specific network structure is shown in figure 2.
The loss function of the artificial deep neural network is calculated by adding the following two loss functions:
S=a*L 1 +b*L 2 wherein L is 1 Representing a network intermediate loss function which is the error between a process value and a real value; l is a radical of an alcohol 2 Representing the final loss function of the network, wherein the final loss function is the error between the output result and the true value, and a and b are weights respectively;
network intermediate loss function L 1 Net final loss function L 2 The following root mean square error loss functions were used:
and S23, obtaining the trained artificial deep neural network through training, and testing the trained artificial deep neural network by using a test set so as to enhance the generalization performance of the artificial deep neural network.
S3, inputting the remote sensing data of the earth surface vegetation in the mining area through a vegetation parameter inversion model, and then outputting the leaf area index, wherein the leaf area index can be displayed in a map plate mode, as shown in FIG. 5. Preferably, the artificial deep neural network further adopts an adam optimization algorithm to obtain higher computational efficiency and lower memory requirements.
The specific embodiments of the invention are as follows: a set of PROSAIL model input parameters of a great mining area of a cylinder are obtained by measuring and viewing data on the spot, and the input parameters cover different vegetation types (shrubs and grasslands) in a research area as wide as possible. Wherein the numerical ranges of the chlorophyll content and the dry matter content of the leaves and the equivalent water thickness meet the truncated Gaussian normal distribution, the numerical ranges given by the leaf area index and the average leaf inclination angle meet the uniform distribution, the soil background spectrum is obtained by adopting a mode of field observation by a spectrometer, and the proportion of dry soil is subjected to the uniform distribution of 0.7-1. The numerical distribution of the solar zenith angle and the observed zenith angle is also set to be uniform distribution, and the average value, the standard deviation, the maximum value and the minimum value of the numerical distribution of each parameter are shown in the following table:
| Landsat | 5 | Landsat 7 | Landsat 8 |
Time of transmission | 1984.3.1 | 1999.4.15 | 2013.2.11 | |
Altitude of satellite | 705km | 705km | 705km | |
Inclination angle of track | 98.2° | 98.2° | 98.2° | |
Period of coverage | 16 days | 16 days | 16 days | |
Size of image | 184km×185.2km | 185km×170km | 170km×180km | |
Sensor mounting | TM | ETM+ | OLI、TIRS | |
Operating conditions | 2011 stop imaging | Run to date | Run to date |
Then, by inputting leaf scale parameters (chlorophyll content, dry matter content, leaf structure parameters, carotene content, brown pigment content, anthocyanin content, equivalent water thickness), canopy scale parameters (leaf area index, average leaf inclination angle), background soil parameters (proportion of dry soil) and observation geometry (sun zenith angle, observation zenith angle and relative azimuth angle) and other information, the following table shows:
obtaining canopy reflectivity data by inputting LAI through forward simulation of a vegetation parameter inversion model; 10000 simulated spectrums are generated randomly in the research, the obtained simulated spectrums are combined with the spectral response functions of Landsat-5, landsat-7 and Landsat-8, and the simulated spectrums are respectively resampled to the wave bands of the corresponding three satellite sensors, so that a vegetation parameter inversion training data set for the three sensors is formed.
The vegetation parameter inversion model in the parameter inversion process is realized on a Google Earth Engine (GEE) platform by adopting an artificial deep neural network optimized based on an ant colony algorithm, the input of the model is the canopy reflectivity of green light, red light, near infrared and short wave infrared bands, for Landsat-5 and Landsat-7, the specific input is the reflectivity of B2, B3, B4, B5 and B7 bands, and for Landsat-8, the input is the reflectivity of B3, B4, B5, B6 and B7 bands. The output of the model is a leaf area index year-by-year synthesized image, and the synthesis algorithm adopts maximum synthesis, namely, the mean value of vegetation parameters of all Landsat effective cloud-free observation inversions of the pixel in summer every year is used as the synthesis value of the final vegetation parameter of the pixel. The preferred synthesis time range during synthesis is July; if July observation can not cover the whole research area, august observation is added during synthesis; if the whole study area can not be covered, the synthesis is carried out by further adding observation of June.
In some embodiments, the remote sensing data of the satellite sensor is derived from Landsat series satellite sensors or Sentinael-2A satellite sensors; the remote sensing data of the satellite sensor comprises six wave bands B1, B2, B3, B4, B5 and B7, the directional reflectivity of the canopy is obtained according to the PROSAIL coupling model, and the directional reflectivity is input into the vegetation parameter inversion model in combination with the spectral response function of the remote sensing data of the satellite sensor for training.
In some embodiments, the remote sensing data of the satellite sensor is derived from a Landsat series satellite sensor, the Landsat series satellite sensor comprises Landsat5, landsat7 and Landsat8, and the remote sensing data of the satellite sensor is registered with the Landsat8 as a reference according to the following method:
selecting a small research area, extracting pixel values of the research area to points, respectively constructing regression linear models among corresponding pixels of different images for Landsat5, landsat7 and Landsat8 by using the least square principle, and realizing registration with Landsat8 as a reference based on the regression linear models. Combining the obtained canopy directional reflectivity data with a spectral response function of Landsat _5-7-8, downloading the spectral response functions of three satellites on a European meteorological satellite organization (EUMETSAT), manually converting the spectral response functions into a txt format file, and obtaining spectral values of corresponding wavelengths of different sensors by an integration method, wherein 10000 analog spectrums are randomly generated in the invention and stored in a csv file, for Landsat-5 and Landsat-7, an analog data set mainly comprises six wave bands B1, B2, B3, B4, B5 and B7 and simulated LAI values (namely leaf area indexes) of corresponding wavelengths, and an analog data set in Landsat-8 mainly comprises six wave bands B2, B3, B4, B5, B6 and B7 and simulated LAI values of corresponding wavelengths; and then, combining the spectral response functions of Landsat-5, landsat-7 and Landsat-8, and resampling the spectral response functions to wave bands corresponding to the three satellite sensors respectively, thereby forming a vegetation parameter inversion training data set aiming at the three sensors. The time span of the invention is 1990-2020, the selected satellite is Landsat-5/7/8 series satellite, and the time resolution is 16 days. The Landsat5 satellite is retired in 2013 in 6 months, the Landsat7 satellite is lifted off in 1999 in 4 months and 5 days in 1999, and Landsat-7ETM + airborne Scanning Line Corrector (SLC) has a fault in 2003 and 31 days in 5 months, so that data stripes of images acquired later are lost, and the use of Landsat ETM remote sensing images is seriously influenced; the Landsat8 satellite is launched to lift off in 2013, 2, month and 11 days, so that the three satellites are selected for monitoring a long time span in a mining area. Because three satellite data are used in the observation process, in order to improve the accuracy of the result, the invention provides a method for calibrating the calculation result by adopting a sensor registration method until Landsat8 is used as a reference. Firstly, selecting the day images of Landsat5, landsat7 and Landsat8 with the closest dates, and respectively calculating the leaf area indexes; then selecting a small research area, extracting pixel values in the area to points, and respectively constructing linear models among pixels corresponding to different images by applying a least square principle; and realizing registration among different sensors based on a regression linear model, and finally taking Landsat8 as a reference. And after the consistency correction of the multi-source result is finished, checking with ground measured data.
An electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the steps of the method for calculating remote sensing of mine area vegetation leaf area index of the present invention.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (8)
1. A remote sensing calculation method for the area index of vegetation leaves on the surface of a mining area is characterized by comprising the following steps: the method comprises the following steps:
s1, coupling a PROSAIL coupling model through a PROSPECT model and an SAILH model, and establishing a vegetation parameter inversion model based on a full-connection network model by adopting the PROSAIL coupling model to couple remote sensing data of a satellite sensor and combining ground actual measurement spectrum and parameter data; input parameters of the PROSAIL coupling model comprise blade dimension, canopy dimension, background soil and observation geometry; the vegetation parameter inversion model comprises a forward simulation model and a reverse simulation model, the forward simulation model realizes a forward simulation process from inputting the leaf area index to outputting the directional reflectivity of the canopy, the reverse simulation model takes the directional reflectivity of the canopy as input and takes the prior leaf area index as output, and the vegetation parameter inversion model continuously adjusts and optimizes the model parameters according to the difference between the minimum estimated leaf area index and the prior leaf area index through the training process of the forward simulation model and the reverse simulation model;
s2, the vegetation parameter inversion model produces a long-time sequence vegetation parameter product on a Google Earth Engine platform, an artificial deep neural network containing the vegetation parameter inversion model is built, the artificial deep neural network adopts an ant colony algorithm, takes the mean square error as the adaptability value of ants, and takes the shortest path determined by ant colonies as the optimal initial weight and bias parameters to be given to the artificial deep neural network for training and testing;
the artificial deep neural network training process is as follows:
s21, carrying out standardization processing on simulation data obtained by the vegetation parameter inversion model and converting the simulation data into dimensionless data; the normalization process uses the following formula:
wherein x represents the normalized value, u represents the mean, and σ represents the standard deviation;
s22, randomly dividing the data after the standardization treatment into a training set and a testing set according to the ratio of 7:3, and training the artificial deep neural network by using the training set to obtain a topological structure of the artificial deep neural network; the loss function of the artificial deep neural network is calculated by adding the following two loss functions:
S=a*L 1 +b*L 2 wherein L is 1 Representing a network intermediate loss function which is an error between a process value and a true value; l is 2 Representing the final loss function of the network, wherein the final loss function is the error between the output result and the true value, and a and b are weights respectively;
network intermediate loss function L 1 Net final loss function L 2 The following root mean square error loss functions were used:
s23, obtaining the trained artificial deep neural network through training, and testing the trained artificial deep neural network by using a test set;
and S3, inputting the mining area earth surface vegetation remote sensing data through a vegetation parameter inversion model, and then outputting a leaf area index.
2. The remote sensing calculation method for the vegetation leaf area index on the ground surface of the mining area according to claim 1, which is characterized by comprising the following steps: the input data of the PROSAIL coupling model comprises leaf scale, canopy scale, background soil and observation geometry, wherein the leaf scale comprises chlorophyll content, dry matter content, leaf structure parameters, carotenoid content, brown pigment content, anthocyanin content and leaf equivalent water thickness, the chlorophyll content, the dry matter content, the leaf equivalent water thickness, the carotenoid content, the brown pigment content and the anthocyanin content are measured by sampling in a mining area range, initial parameters of the carotenoid content, the brown pigment content and the anthocyanin content are set to be zero, and the range of leaf structure parameter values is set to be between 1 and 2; the canopy scale comprises a leaf area index and an average leaf inclination angle, and the leaf area index and the average leaf inclination angle are obtained through field measurement; the background soil comprises the proportion of dry soil and soil spectrum data, the proportion of the dry soil is measured by sampling in a mining area, and the soil spectrum data is obtained by field measurement; the observation geometry comprises a solar zenith angle, an observation zenith angle and a relative azimuth angle, and the observation geometry is derived from remote sensing data of a satellite sensor; the chlorophyll content, the dry matter content, the leaf equivalent water thickness, the carotenoid content, the brown pigment content, the anthocyanin content, the proportion of dry soil and the soil spectrum data are measured for multiple times to determine the numerical distribution, the average value, the maximum value and the minimum value of the parameters.
3. The mining area earth surface vegetation leaf area index remote sensing calculation method according to claim 1, characterized by: the remote sensing data of the satellite sensor is from Landsat series satellite sensors or Sentiniel-2A satellite sensors; the remote sensing data of the satellite sensor comprises six wave bands B1, B2, B3, B4, B5 and B7, the directional reflectivity of the canopy is obtained according to the PROSAIL coupling model, and the directional reflectivity is input into the vegetation parameter inversion model in combination with the spectral response function of the remote sensing data of the satellite sensor for training.
4. The mining area earth surface vegetation leaf area index remote sensing calculation method according to claim 1, characterized by: the remote sensing data of the satellite sensor is from Landsat series satellite sensors, the Landsat series satellite sensors comprise Landsat5, landsat7 and Landsat8, and the remote sensing data of the satellite sensor is subjected to registration processing by taking Landsat8 as a reference according to the following method:
selecting a small research area, extracting pixel values of the research area to points, respectively constructing regression linear models among corresponding pixels of different images for Landsat5, landsat7 and Landsat8 by using the least square principle, and realizing registration with Landsat8 as a reference based on the regression linear models.
5. The mining area earth surface vegetation leaf area index remote sensing calculation method according to claim 1, characterized by: and the artificial deep neural network calculates a loss function according to the leaf area index obtained by simulation and the leaf area index obtained by actual measurement, and optimizes the weight and the bias parameter by using the loss value and the optimization function.
6. The mining area earth surface vegetation leaf area index remote sensing calculation method according to claim 1, characterized by: in the method S22, the artificial deep neural network further adopts an adam optimization algorithm to obtain higher computational efficiency and lower memory requirement.
7. The remote sensing calculation method for the vegetation leaf area index on the ground surface of the mining area according to claim 1, which is characterized by comprising the following steps: the artificial deep neural network comprises an input layer, a hidden layer and an output layer, wherein the input layer is used for inputting a vegetation parameter inversion training data set, the hidden layer is used for ant colony algorithm processing, and the output layer is used for outputting a leaf area index.
8. An electronic device, characterized in that: the method comprises the following steps: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211120240.6A CN115526098B (en) | 2022-09-14 | 2022-09-14 | Remote sensing calculation method for leaf area index of surface vegetation in mining area and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211120240.6A CN115526098B (en) | 2022-09-14 | 2022-09-14 | Remote sensing calculation method for leaf area index of surface vegetation in mining area and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115526098A CN115526098A (en) | 2022-12-27 |
CN115526098B true CN115526098B (en) | 2023-04-18 |
Family
ID=84697156
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211120240.6A Active CN115526098B (en) | 2022-09-14 | 2022-09-14 | Remote sensing calculation method for leaf area index of surface vegetation in mining area and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115526098B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116503747B (en) * | 2023-06-30 | 2023-09-15 | 华中农业大学 | Heterogeneous earth surface leaf area index inversion method based on multi-scale remote sensing |
CN117408418B (en) * | 2023-10-16 | 2024-05-10 | 河南大学 | Regional comprehensive ecological environment quality evaluation method and system based on remote sensing big data |
CN117493733B (en) * | 2023-10-30 | 2024-06-11 | 武汉大学 | Method and system for calculating total primary productivity based on satellite multispectral information |
CN118039163B (en) * | 2024-04-07 | 2024-06-25 | 北京惠每云科技有限公司 | Clinical risk management method and device based on large model and ant colony algorithm |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112595267A (en) * | 2020-12-09 | 2021-04-02 | 中国科学院空天信息创新研究院 | Summer corn LAI inversion method fusing factor analysis and ant colony wavelet neural network model |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111598243A (en) * | 2020-04-30 | 2020-08-28 | 中国科学院空天信息创新研究院 | Leaf area index inversion model based on SAR and WFV data |
CN113128134B (en) * | 2021-06-17 | 2021-09-14 | 中国矿业大学(北京) | Mining area ecological environment evolution driving factor weight quantitative analysis method |
CN113919227B (en) * | 2021-10-15 | 2022-04-15 | 中国矿业大学(北京) | Mining area ecological time accumulation effect point and space accumulation range identification method |
CN114021348B (en) * | 2021-11-05 | 2022-05-20 | 中国矿业大学(北京) | Fine land utilization type mining area vegetation carbon sink remote sensing inversion method |
-
2022
- 2022-09-14 CN CN202211120240.6A patent/CN115526098B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112595267A (en) * | 2020-12-09 | 2021-04-02 | 中国科学院空天信息创新研究院 | Summer corn LAI inversion method fusing factor analysis and ant colony wavelet neural network model |
Also Published As
Publication number | Publication date |
---|---|
CN115526098A (en) | 2022-12-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN115526098B (en) | Remote sensing calculation method for leaf area index of surface vegetation in mining area and electronic equipment | |
CN110222475B (en) | Method for inverting moisture content of winter wheat plants based on multispectral remote sensing of unmanned aerial vehicle | |
Yu et al. | Improvement of sugarcane yield estimation by assimilating UAV-derived plant height observations | |
Hu et al. | Improvement of sugarcane crop simulation by SWAP-WOFOST model via data assimilation | |
CN108985588A (en) | A kind of crop yield prediction remote sensing estimation method, device and system | |
CN110751094A (en) | Crop yield estimation technology based on GEE comprehensive remote sensing image and deep learning method | |
CN113268923A (en) | Summer corn yield estimation method based on simulation multispectral | |
CN114510824B (en) | Construction method of regional scale evapotranspiration model synchronously considering dynamic changes of vegetation canopy and root system | |
CN115479960B (en) | Method for monitoring soil humidity in crop growth process by combining SAR (synthetic aperture radar) with optical remote sensing data | |
CN108106676A (en) | A kind of monitoring method and device of the crops Spring frost based on remotely-sensed data | |
Song et al. | Estimating effective leaf area index of winter wheat using simulated observation on unmanned aerial vehicle-based point cloud data | |
Deb et al. | An ensemble data assimilation approach to improve farm-scale actual evapotranspiration estimation | |
Zhang et al. | Evaluating maize evapotranspiration using high-resolution UAV-based imagery and FAO-56 dual crop coefficient approach | |
CN115439402A (en) | Sugar cane yield-per-unit remote sensing estimation method and device based on SAFY-Sugar model | |
Zhai et al. | CatBoost algorithm for estimating maize above-ground biomass using unmanned aerial vehicle-based multi-source sensor data and SPAD values | |
Zhangzhong et al. | Development of an evapotranspiration estimation method for lettuce via mobile phones using machine vision: Proof of concept | |
Zhang et al. | Estimation of rice yield from a C-band radar remote sensing image by integrating a physical scattering model and an optimization algorithm | |
Li et al. | Daily monitoring of Effective Green Area Index and Vegetation Chlorophyll Content from continuous acquisitions of a multi-band spectrometer over winter wheat | |
US20240167947A1 (en) | CEEMDAN-based method for screening and monitoring soil moisture stress in agricultural fields | |
CN117011694A (en) | Forest tree growth parameter prediction method based on cascade circulation network | |
CN112949182B (en) | Method for calculating regional scale maximum pore conductance of coupling surface soil water content | |
Tang et al. | UAV-based high spatial and temporal resolution monitoring and mapping of surface moisture status in a vineyard | |
CN112380497B (en) | Straw coefficient estimation method and system for regional corn | |
CN115994939A (en) | Tree leaf area estimation method based on ground laser point cloud | |
CN115223062A (en) | UAV data-based method for correcting forest stand accumulation amount time difference of eucalyptus artificial forest region |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |