CN111738974B - Leaf area index ground sampling method - Google Patents
Leaf area index ground sampling method Download PDFInfo
- Publication number
- CN111738974B CN111738974B CN201910231598.8A CN201910231598A CN111738974B CN 111738974 B CN111738974 B CN 111738974B CN 201910231598 A CN201910231598 A CN 201910231598A CN 111738974 B CN111738974 B CN 111738974B
- Authority
- CN
- China
- Prior art keywords
- leaf area
- index
- map
- normalized difference
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Abstract
The invention provides a leaf area index ground sampling method, which comprises the following steps: acquiring a remote sensing image; obtaining a normalized difference vegetation index map and a leaf area index map; obtaining a polymerized leaf area index map and a polymerized normalized difference vegetation index map; determining a correlation index map by utilizing a fitting relation between leaf area indexes and vegetation indexes and an aggregate normalized vegetation index variance map, and selecting the minimum E correlation indexes; determining a simulated leaf area index map by calculating a geometric model, E related indexes, a normalized difference vegetation index map, a leaf area index map and an aggregate normalized difference vegetation index map; and calculating a deviation value of the simulated leaf area index graph relative to the aggregate leaf area index graph, and determining a target sampling area according to the deviation value. The invention relieves the problems that the sampling process of the existing method is low in efficiency and cannot meet the actual application requirements, and achieves the effects of improving the sampling process efficiency and meeting the actual application requirements.
Description
Technical Field
The invention relates to the technical field of ground sampling, in particular to a leaf area index ground sampling method.
Background
Leaf Area Index (LAI) is defined as the sum of Leaf areas of a single-sided green Leaf per unit Area of a plant and is a key parameter in characterizing the growth of vegetation. The remote sensing technology can rapidly acquire the large-area surface leaf area index, and in order to verify the accuracy of the remote sensing inversion leaf area index product, ground truth values are generally acquired as comprehensively as possible to carry out inspection analysis on the inversion product. Because of the manpower and time constraints, limited ground truth measurement cannot effectively meet the authenticity inspection requirements of leaf area index products, and therefore, it is desirable to design a ground sampling strategy with as few ground sampling points as possible to represent as comprehensive ground truth as possible.
The uniform sampling method, the random sampling method, the sampling method based on priori knowledge and the like are the most commonly used leaf area index ground sampling method at present. The uniform sampling method adopts a regular sampling mode to acquire ground truth values of the whole area, and sampling points are uniformly distributed in space and have good discreteness; the random sampling method adopts a random mode to acquire ground sample points, is simple and direct, and has balanced probability of all the sample points in the whole area; the sampling method based on priori knowledge reduces uncertainty of sample point information through collection and analysis of priori data, and effectively improves representativeness of ground measured values to the whole area. However, due to lack of adequate knowledge of heterogeneous earth surfaces, these sampling methods can meet the leaf area index remote sensing product verification requirements of uniform earth surfaces, but are difficult to compete with areas of higher non-uniformity. In practical application, when the uniform sampling method is applied to irregular heterogeneous ground surfaces, the problems that some vegetation types are not sampled enough or even not collected can occur; the random sampling method is easy to cause the problems of sample overlapping and weak sample representativeness; when the priori data is insufficient, the sampling method based on the priori knowledge is easy to have the problem of insufficient samples, so that the sampling process is low in efficiency and cannot meet the actual application requirements.
Disclosure of Invention
First, the technical problem to be solved
In view of the above technical problems, the present invention aims to provide a method for sampling a leaf area index ground, so as to alleviate the problems that the efficiency of the heterogeneous earth surface leaf area index ground sampling process in the prior art is low and the actual application requirements cannot be satisfied.
(II) technical scheme
According to one aspect of the present invention, there is provided a leaf area index ground sampling method comprising:
step S1: acquiring a remote sensing image of a region to be sampled, and preprocessing the remote sensing image;
step S2: calculating a normalized difference vegetation index map and a leaf area index map by using the preprocessed remote sensing image;
step S3: carrying out average polymerization treatment on the leaf area index map to obtain a polymerized leaf area index map;
step S4: performing aggregation treatment on the normalized difference vegetation index map to obtain an aggregation normalized difference vegetation index map and an aggregation normalized difference vegetation index variance map; determining a correlation index map by utilizing the aggregate normalized difference vegetation index variance map, and selecting the minimum E correlation indexes;
step S5: calculating a simulated leaf area index map from the normalized difference vegetation index map, the leaf area index map, the aggregate normalized difference vegetation index and the E correlation indexes;
Step S6: calculating a deviation value of the simulated leaf area index graph relative to the aggregate leaf area index graph, and determining a target sampling area according to the deviation value.
In some embodiments, the S1 specifically includes:
s1-1: acquiring a remote sensing image of a region to be sampled, wherein the image ground object comprises vegetation type ground objects, and the image wave band comprises a red wave band and a near infrared wave band;
s1-2: performing radiometric calibration on the remote sensing image, and converting the gray value of each pixel in the remote sensing image into a radiation brightness value;
s1-3: performing atmospheric correction on the remote sensing image, and determining the reflectivity value of each pixel in the remote sensing image;
s1-4: and performing geometric correction on the remote sensing image, and determining longitude and latitude information of each pixel in the remote sensing image and the ground area corresponding to each pixel.
In some embodiments, the S2 specifically includes:
s2-1: for the remote sensing image, calculating to obtain a normalized difference vegetation index map by utilizing the reflectivity R of a red wave band and the reflectivity NIR of a near infrared wave band;
wherein NDVI represents a normalized difference vegetation index;
s2-2: determining a fitting relation between the leaf area index and the vegetation index;
s2-3: and calculating a leaf area index map by using the fitting relation between the leaf area index and the vegetation index and the normalized difference vegetation index map.
In some embodiments, the determining a fit relationship of the leaf area index to the vegetation index comprises:
simulating to obtain a plurality of simulated leaf area indexes and simulated vegetation spectrums respectively corresponding to the simulated leaf area indexes by using preset vegetation parameter information and a radiation transmission model;
calculating a simulated normalized difference vegetation index by using the simulated vegetation spectrum for each simulated vegetation spectrum;
substituting the simulated leaf area index and the corresponding simulated normalized difference vegetation index into a preset formula for each simulated leaf area index, determining a fitting coefficient of the preset formula, and obtaining a fitting relation between the leaf area index and the vegetation index, wherein the preset formula is as follows:
wherein, LAI i For the ith simulated leaf area index, NDVI i Normalizing the differential vegetation index for the ith simulation, and a and b are the fitting coefficients.
In some embodiments, the S3 specifically includes:
s3-1: setting N rows and M columns of aggregation windows, and moving the aggregation windows in the horizontal direction from left to right, wherein the pixels covered on the leaf area index map of the aggregation windows are not overlapped each time; calculating an average value of N multiplied by M leaf area index values in the aggregation window to obtain the aggregation leaf area index;
S3-2: aggregating pixel values of rows 1 to N on the leaf area index map to obtain a first row leaf area index value of the aggregate leaf area index map; and polymerizing pixels of the (n+1) -2N rows on the leaf area index map to obtain leaf area index values of a second row of the polymerized leaf area index map until the last row is polymerized to obtain the polymerized leaf area index map.
In some embodiments, the S4 specifically includes:
s4-1: performing aggregation treatment on the normalized difference vegetation index map to obtain an aggregation normalized difference vegetation index map and an aggregation normalized difference vegetation index variance map;
s4-2: determining a related index map by using a fitting relation between the leaf area index and the vegetation index and the aggregate normalized difference vegetation index variance map;
s4-3: the smallest E correlation indexes are chosen.
In some embodiments, the aggregating the normalized differential vegetation index map specifically includes:
setting N rows and M columns of aggregation windows, and moving the aggregation windows in the horizontal direction from left to right, wherein pixels covered on the normalized difference vegetation index map are not overlapped each time by the aggregation windows; solving an average value of N multiplied by M pixel values in the aggregation window to obtain the aggregation normalized difference vegetation index; solving variances for N multiplied by M pixel values in the aggregation window to obtain the aggregation normalized difference vegetation index variances;
On the normalized difference vegetation index map, the pixel value aggregation processing results of the 1 st to N th rows are pixel values of the aggregation normalized difference vegetation index map and the first row of the aggregation normalized difference vegetation index variance map;
and aggregating pixels of the (n+1) -th row to the (2N) -th row on the image to obtain pixel values of a second row of the aggregate normalized difference vegetation index map and the aggregate normalized difference vegetation index variance map until the last row is aggregated to obtain the aggregate normalized difference vegetation index map and the aggregate normalized difference vegetation index variance map.
In some embodiments, the correlation index map determining process specifically includes:
calculating a correlation index of each pixel by using the aggregate normalized difference vegetation index variance diagram and a correlation index calculation formula to obtain the correlation index diagram, wherein the correlation index calculation formula is as follows:
wherein f (x) =a×e b*x For the fitting relation between the leaf area index and the vegetation index, a and b are fitting coefficients, N i For the vegetation index value of the ith pixel in the aggregate normalized difference vegetation index map, L1 i For the ith pixel value in the aggregate leaf area index map, f' is the second order Taylor expansion formula of f (x), nvar i CI for variance value of ith pixel in the aggregate normalized difference vegetation index variance map i Is the correlation index of the i-th picture element.
In some embodiments, the selecting process of the E correlation indexes specifically includes:
calculating standard deviation of leaf area indexes corresponding to all pixels in the leaf area index map;
calculating the initial sampling number by using the standard deviation and an initial sampling number calculation formula, wherein the initial sampling number calculation formula is as follows:
and num is the sampling number, the sampling number is equal to the number of related indexes, S is the standard deviation, and Z and d are known parameters.
In some embodiments, the S5 specifically includes:
s5-1: forming a point set to be modeled by a target number of normalized difference vegetation indexes and the leaf area indexes corresponding to the E correlation indexes, wherein the target number is equal to the E multiplied by N, and the obtained result is multiplied by M;
s5-2: calculating an upper convex hull function and a lower convex hull function of the point set to be built by using a preset function, and generating a calculation geometric model by using the upper convex hull function and the lower convex hull function; the computational geometry model is:
L2 i ＝[g^(N i )+g v (N i )]/2
wherein L2 i For the ith pixel value, N in the simulated leaf area index map i For the ith pixel value in the aggregate normalized difference vegetation index map, g≡x is the upper convex hull function, g v (x) Is the lower convex hull function;
s5-3: and sequentially inputting each pixel value in the aggregation normalized difference vegetation index map into the calculation geometric model, and calculating to obtain the simulated leaf area index map.
In some embodiments, the step S6 specifically includes:
s6-1: performing deviation calculation on the simulated leaf area index map and the aggregate leaf area index map to obtain a deviation value;
wherein Bias is the offset value, L2 i L1 is the ith pixel value in the simulated leaf area index map i The ith pixel value in the aggregate leaf area index chart is used, and n is the total pixel number;
s6-2: judging a target sampling area according to preset conditions, wherein the preset conditions are as follows:
and if the deviation value meets a preset condition, determining the ground area corresponding to the E correlation indexes in the leaf area index graph as the target sampling area, otherwise, adding 1 on the basis of the initial sampling number and executing step S5.
(III) beneficial effects
From the above technical solution, the ground sampling method provided by the invention has the following beneficial effects:
(1) The invention provides a leaf area index ground sampling method, which utilizes a correlation index to select a high-representative target sampling area, so that the problems of insufficient sampling, overlapping samples and weak sample representativeness can not occur, the problem of insufficient samples caused by insufficient prior data can not occur, and the technical effects of improving the efficiency of a sampling process and meeting the actual application requirements can be achieved;
(2) The leaf area index ground sampling method provided by the invention considers the influence of ground heterogeneity and does not use priori knowledge, and a target sampling area is directly determined on the remote sensing image of the area to be sampled, so that the ground sampling method provided by the invention is simple and efficient, can be directly applied to actually measured collection of leaf area indexes of heterogeneous ground surfaces, and can quickly generate an effective ground sampling scheme under the condition of insufficient prior knowledge; meanwhile, the ground sampling method provided by the invention is applicable to any vegetation type, and is universal, convenient and easy to popularize and use.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a leaf area index ground sampling method provided by an embodiment of the invention;
FIG. 2 is a schematic diagram of a normalized differential vegetation index graph according to an embodiment of the present invention;
FIG. 3 is a schematic view of a leaf area index graph provided by an embodiment of the present invention;
FIG. 4 is a schematic diagram of a polymer leaf area index graph provided by an embodiment of the present invention;
FIG. 5 is a schematic diagram of an aggregate normalized difference vegetation index graph according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of an aggregate normalized difference vegetation index variance diagram according to an embodiment of the present invention;
FIG. 7 is a schematic diagram of a normalized difference vegetation index map after correlation index selection according to an embodiment of the present invention;
FIG. 8 is a graph of leaf area index after correlation index selection according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The leaf area index ground sampling method provided by the embodiment of the invention can be used for relieving the technical problems that the efficiency of the sampling process is low and the actual application requirements cannot be met in the prior art, and achieves the technical effects of improving the efficiency of the sampling process and meeting the actual application requirements.
For the sake of understanding the present embodiment, a detailed description will be given first of a method for sampling a leaf area index ground according to an embodiment of the present invention, and as shown in fig. 1 to 8, the leaf area index ground sampling method may include the following steps.
Step S1: and acquiring a remote sensing image of the region to be sampled, and preprocessing the remote sensing image. The detailed steps are as follows:
(1) The remote sensing image of the area to be sampled contains vegetation type ground objects, the wave bands must include red wave bands and near infrared wave bands, and the ground surface of the area to be sampled can be a heterogeneous ground surface or a uniform ground surface by way of example.
(2) The remote sensing image preprocessing comprises radiometric calibration, atmospheric correction and geometric correction, and comprises the following specific steps:
firstly, performing radiometric calibration on a remote sensing image by using a radiometric calibration formula, and converting a gray value of each pixel into a radiance value, wherein the radiometric calibration formula is as follows:
L i ＝DN i *gain+offset
Wherein L is i For the radiation brightness value, DN, of the ith pixel in the remote sensing image i For the gray value of the i-th pixel in the remote sensing image, gain is a known gain value, offset is a known offset value, and the gain value and the offset value can be obtained from the satellite remote sensing product description.
And then, carrying out atmosphere correction on The remote sensing image, and determining The reflectivity value of each pixel in The remote sensing image, wherein The remote sensing image can be subjected to atmosphere correction on The basis of a FLAASH (The Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes) module by utilizing ENVI (The ENvironment for Visualizing Images) software.
And finally, performing geometric correction on the remote sensing image, and determining the ground area corresponding to each pixel in the remote sensing image, wherein geometric projection information can be attached to the remote sensing image by using ENVI software and based on an image to image mode, or by using a ground control point and ENVI image to map method.
Step S2: and calculating a normalized difference vegetation index map and a leaf area index map by using the remote sensing image, wherein the normalized difference vegetation index map and the leaf area index map are shown in fig. 2. The detailed steps are as follows:
(1) And calculating to obtain a normalized difference vegetation index (Normalized Difference Vegetation Index, NDVI) by utilizing the red band reflectivity and the near infrared band reflectivity in the preprocessed remote sensing image.
Wherein, NDVI i Normalized difference vegetation index, NIR, for the ith pixel in the remote sensing image i Is the first in the remote sensing imageNear infrared band reflectivity of i pixels, R i And the reflectivity of the red wave band of the ith pixel in the remote sensing image is obtained.
(2) The fitting relation between the leaf area index and the vegetation index is determined, and the method comprises the following specific steps:
firstly, a plurality of simulated leaf area indexes and simulated vegetation spectrums respectively corresponding to the simulated leaf area indexes are obtained through simulation by using preset vegetation parameter information and a radiation transmission model.
For example, the vegetation parameter information may be different vegetation parameter information provided by the LOPEX'93 (Leaf Optical Properties Experiment) database. The radiation delivery model may be PROSAIL (PROSPECT+SAIL: properties spectra + scattering by arbitrarily inclined leaves).
And then, carrying out convolution processing on each simulated vegetation spectrum by utilizing a spectral response function to obtain the simulated red wave band reflectivity and the simulated near infrared reflectivity.
Wherein R is i And NIR (near infrared) i The ith simulated red band reflectivity and the simulated near infrared reflectivity, f R (lambda) and f NIR (lambda) is the known red band spectral response function and the known near infrared band spectral response function respectively, and can be obtained from satellite remote sensing product information, lambda is the wavelength, ρ i (lambda) is the simulated vegetation spectrum at wavelength lambda.
And then, obtaining a simulated normalized difference vegetation index by using the NDVI calculation formula, the simulated red wave band reflectivity and the simulated near infrared reflectivity.
Wherein, NDVI i Normalizing the differential vegetation index for the ith simulation, NIR i For the ith simulated near infrared band reflectivity, R i The i-th analog red wave Duan Fanshe rate.
And substituting the simulated leaf area index and the corresponding simulated normalized difference vegetation index into a preset formula, and determining a fitting coefficient of the preset formula to obtain a fitting relation between the leaf area index and the vegetation index.
The preset formula may be:
in a preset formula, LAI i For the ith simulated leaf area index, NDVI i The differential vegetation index is normalized for the ith simulation, and a and b are fitting coefficients. When a and b are determined, the preset formula is a fitting relation formula of the vegetation index and the leaf area index.
And finally, calculating to obtain a leaf area index by using the normalized difference vegetation index and the fitting relation.
In embodiments of the present invention, the normalized difference vegetation index may be replaced with a modified normalized difference vegetation index (Modified Normalized Difference Vegetation Index, MNDVI), a Ratio vegetation index (SR), or a vertical vegetation index (Perpendiculars Vegetation Index, PVI).
A great deal of researches show that the vegetation index has a linear or nonlinear relation with the leaf area index, the normalized difference vegetation index is the most widely applied one, and the calculation of the vegetation leaf area index of the area to be sampled can be rapidly realized by establishing the empirical relation between the leaf area index and the normalized difference vegetation index.
The leaf area index can be calculated by using a simple regression modeling generation empirical formula through ground measured data, and can be inverted by using a physical method based on a radiation transmission model.
Step S3: and carrying out average polymerization treatment on the leaf area index graph to obtain a polymerized leaf area index graph. The detailed steps are as follows:
(1) And setting N rows and M columns of aggregation windows, moving the aggregation windows in the horizontal direction from left to right, wherein pixels covered on the leaf area index graph by the aggregation windows are not overlapped each time, and solving an average value of N multiplied by M leaf area index values in the aggregation windows to obtain the aggregation leaf area index.
Wherein the aggregate leaf area index may be calculated using a formula, which may be:
in this formula, L1 k For the polymer leaf area index obtained for the kth polymerization,for the leaf area index corresponding to the kth pixel covered by the aggregation window in the leaf area index map,/I>The position in the aggregation window is the ith row and the jth column. The values of N and M can be set according to actual needs. In the embodiment of the present invention, n=m=3 is taken as an example for explanation.
(2) And aggregating pixel values of the 1 st to N th rows on the leaf area index map to obtain a first row leaf area index of the aggregate leaf area index map.
And if the row number of the aggregation window is 3, aggregating pixels of the 1 st to 3 rd rows on the leaf area index graph. On the leaf area index chart shown in FIG. 3, the 3 rows and 3 columns of aggregation windows are first covered with leaf area indexes L11, L12, L13, L21, L22, L23, L31, L32 and L33 corresponding to 9 pixels to obtain a first aggregate leaf area index L1 1 Wherein, the method comprises the steps of, wherein,
then, the aggregation windows move rightward on the 1 st to 3 rd rows of the leaf area index map, and the 3 rd rows and 3 th columns of aggregation windows cover the leaf area indexes L14, L15, L16, L24, L25, L26, L34, L35 and L36 corresponding to 9 pixels to obtain a second aggregation leaf area index L1 2 Wherein, the method comprises the steps of, wherein,
then, the aggregation windows move rightward on the 1 st to 3 rd rows of the leaf area index map, and the 3 rd rows and 3 columns of aggregation windows cover the leaf area indexes L17, L18, L19, L27, L28, L29, L37, L38 and L39 corresponding to 9 pixels to obtain a third aggregation leaf area index L1 3 Wherein, the method comprises the steps of, wherein,
here, the pixels of rows 1 to 3 of the leaf area index map are aggregated. The aggregation window then continues to aggregate from the pixels of rows 4 to 6 of the leaf area index map.
And polymerizing pixels of the (n+1) -2N rows on the leaf area index map to obtain leaf area index values of a second row of the polymerized leaf area index map until the last row is polymerized to obtain the polymerized leaf area index map.
Wherein, the aggregation window is on the 4 th to 6 th rows of the leaf area index diagram, the aggregation window of 3 rows and 3 columns firstly covers the leaf area indexes L41, L42, L43, L51, L52, L53, L61, L62 and L63 corresponding to 9 pixels to obtain a fourth aggregation leaf area index L1 4 Wherein, the method comprises the steps of, wherein,
then, the aggregation window is upward on lines 4 to 6 of the leaf area index mapMoving right, the 3-row 3-column aggregation window covers leaf area indexes L44, L45, L46, L54, L55, L56, L64, L65 and L66 corresponding to 9 pixels to obtain a fifth aggregation leaf area index L1 5 Wherein, the method comprises the steps of, wherein,
then, the aggregation windows move rightward on the 4 th to 6 th rows of the leaf area index map, and the aggregation windows of 3 rows and 3 columns cover the leaf area indexes L47, L48, L49, L57, L58, L59, L67, L68 and L69 corresponding to 9 pixels to obtain a sixth aggregation leaf area index L1 6 Wherein, the method comprises the steps of, wherein,
here, the pixels of rows 4 to 6 of the leaf area index map are aggregated. The aggregation window then continues to aggregate from the pixels of rows 7 to 9 of the leaf area index map.
Wherein, the aggregation window of 3 rows and 3 columns firstly covers leaf area indexes L71, L72, L73, L81, L82, L83, L91, L92 and L93 corresponding to 9 pixels to obtain a seventh aggregation leaf area index L1 7 Wherein, the method comprises the steps of, wherein,
then, the aggregation windows move rightward on the 7 th to 9 th rows of the leaf area index map, and the aggregation windows of 3 rows and 3 columns cover the leaf area indexes L74, L75, L76, L84, L85, L86, L94, L95 and L96 corresponding to 9 pixels to obtain an eighth aggregation leaf area index L1 8 Wherein, the method comprises the steps of, wherein,
then, the aggregation window is at lines 7 to 9 of the leaf area index mapMoving upward to the right, the aggregation window of 3 rows and 3 columns covers leaf area indexes L77, L78, L79, L87, L88, L89, L97, L98 and L99 corresponding to 9 pixels to obtain a ninth aggregation leaf area index L1 9 Wherein, the method comprises the steps of, wherein,
here, the pixels of rows 7 to 9 of the leaf area index map are aggregated. That is, the leaf area index map is aggregated into an aggregate leaf area index map, which may be as shown in fig. 4.
Step S4: carrying out average aggregation treatment on the normalized difference vegetation index map to obtain an aggregation normalized difference vegetation index map and an aggregation normalized difference vegetation index variance map; and determining a correlation index map by using a fitting relation between the leaf area index and the vegetation index and the aggregate normalized vegetation index variance map, and selecting the minimum E correlation indexes. The detailed steps are as follows:
(1) And carrying out aggregation treatment on the normalized difference vegetation index map to obtain an aggregation normalized difference vegetation index map and an aggregation normalized difference vegetation index variance map. The method comprises the following specific steps:
and the aggregation window moves in the horizontal direction from left to right, pixels covered on the normalized difference vegetation index graph by the aggregation window are not overlapped each time, the average value is obtained for N multiplied by M normalized difference vegetation index values in the aggregation window to obtain the aggregation normalized difference vegetation index, and the variance is obtained for N multiplied by M normalized difference vegetation index values in the aggregation window to obtain the aggregation normalized difference vegetation index variance.
Wherein the mean and variance may be calculated using a formula, which may be:
in this formula, nveg k 、Nvar k Respectively obtaining an aggregation normalized difference vegetation index and an aggregation normalized difference vegetation index variance by the kth aggregation,for the k-th normalized difference vegetation index covered by the aggregation window in the normalized difference vegetation index map,/v>The position in the aggregation window is the ith row and the jth column.
And on the normalized difference vegetation index map, the pixels of the 1 st to N rows are polymerized to obtain the polymerized normalized difference vegetation index map and the pixel value of the first row of the polymerized normalized difference vegetation index variance map.
Aggregating pixels of the (n+1) -2N rows on the image to obtain pixel values of a second row of the aggregate normalized difference vegetation index map and the aggregate normalized difference vegetation index variance map until the last row is aggregated to obtain the aggregate normalized difference vegetation index map and the aggregate normalized difference vegetation index variance map;
details of the implementation of the above polymerization process may be referred to as a leaf area index polymerization process.
The aggregate normalized difference vegetation index map and the aggregate normalized difference vegetation index variance map are shown in fig. 5 and 6, respectively.
(2) And determining a related index map by utilizing a fitting relation between the leaf area index and the vegetation index and the aggregate normalized vegetation index variance map, wherein the specific steps are as follows:
calculating a correlation index of each pixel by using the aggregate normalized difference vegetation index variance diagram and a correlation index calculation formula to obtain the correlation index diagram, wherein the correlation index calculation formula is as follows:
CI i ＝f(N i )-L1 i
in this formula, CI i For the correlation index of the i-th pixel in the aggregate leaf area index map, f (x) =a×e b*x N is the relation between the vegetation index and leaf area index i For the ith pixel value, L1, in the aggregate normalized differential vegetation index map i And (3) the i pixel value in the aggregate leaf area index graph. Correlation index CI i The smaller the value of (C) is, the description CI i The better the representativeness of the region to be sampled corresponding to the pixel is, the more should be determined as the target sampling region.
In the embodiment of the invention, the leaf area index ground sampling method can further comprise the following steps:
substituting the aggregate normalized difference vegetation index into a preset correlation index calculation formula to calculate the correlation index of each pixel in the aggregate leaf area index map, wherein the preset correlation index calculation formula is as follows:
Wherein f' is the second order Taylor expansion formula of f (x), nvar i And normalizing the ith pixel value in the differential vegetation index variance map for the aggregation.
In the embodiment of the invention, each pixel in the aggregate leaf area index map corresponds to a relevant index, and the aggregate leaf area index of the first pixel in the aggregate leaf area index map is L1 1 The correlation index of the first pixel is CI 1 . The second pixel in the polymer leaf area index map has a polymer leaf area index L1 2 The correlation index of the second pixel is CI 2 . Up to the ninth pixel in the polymer leaf area index map having a polymer leaf area index of L1 9 The correlation index of the ninth pixel is CI 9 . And sorting the 9 correlation indexes, and selecting the minimum E correlation indexes. In the embodiment of the invention, the sampling number is less than or equal to 9.
(3) The determining of the sampling number E may include the steps of:
and calculating standard deviation of leaf area indexes corresponding to all pixels in the leaf area index map.
The leaf area index map shown in fig. 3 is taken as an example. The standard deviation of the overall data was calculated with L11, L12, & gt, L19, L21, L22, & gt, L29, & gt, and L91, L92, & gt, L99 as the overall data.
Calculating the sampling number E by using the standard deviation and an initial sampling number calculation formula, wherein the initial sampling number calculation formula is as follows:
wherein num is the sampling number E, E is a positive integer, S is the standard deviation, and Z and d are both known parameters. Z represents statistics at a confidence level, e.g., a Z statistic of 1.96 at a 95% confidence level and a Z statistic of 2.68 at a 99% confidence level. Illustratively, in the embodiment of the present invention, z=1.96 may be taken. d is the desired accuracy, and here, the value is set with reference to the verification accuracy (RMSE: 0.73) of the MODIS leaf area index product, and in the embodiment of the present invention, d=0.73 may be set.
Step S5: and determining a simulated leaf area index map by using the normalized difference vegetation index map, the leaf area index map, the aggregate normalized difference vegetation index and E related indexes. The detailed steps are as follows:
and forming a point set to be modeled by the target number of normalized difference vegetation indexes corresponding to the sampled plurality of correlation indexes and the target number of leaf area indexes, wherein the target number is equal to E multiplied by N, and the obtained result is multiplied by M.
In the embodiment of the invention, the sampling number E is equal to 4, and the minimum 4 correlation indexes are selected as CI 1 、CI 5 、CI 6 And CI (CI) 8 An example is described. That is, CI will be 1 、CI 5 、CI 6 And CI (CI) 8 Corresponding 36 (4×3×3=3)36 The normalized differential vegetation indices and 36 (4×3×3=36) leaf area indices make up the point set to be modeled. CI (CI) 1 、CI 5 、CI 6 And CI (CI) 8 The first set of 36 corresponding normalized difference vegetation indices may be as shown in fig. 6. CI (CI) 1 、CI 5 、CI 6 And CI (CI) 8 The second set of corresponding 36 leaf area indices may be as shown in fig. 7. Since the normalized difference vegetation index in the first number corresponds to the leaf area index in the second number, a set of points to be modeled { (N11, L11), (N12, L12), (N13, L13), (N21, L21),., (N95, L95), (N96, L96) } can be obtained. The set of points to be modeled includes 36 points.
And calculating an upper convex hull function and a lower convex hull function of the point set to be modeled by using a preset function.
The preset function may be, for example, a convhulln function of matlab.
And generating a calculation geometric model by using the upper convex hull function and the lower convex hull function.
And sequentially inputting each pixel value in the aggregation normalized difference vegetation index map into the calculation geometric model, and calculating to obtain the simulated leaf area index map.
Wherein, the calculation geometric model can be:
L2 i ＝[g^(N i )+g v (N i )]/2
In the geometric model, L2 i For the simulated leaf area index of the ith pixel in the simulated leaf area index map, g≡x is the upper convex hull function, g v (x) For the lower convex hull function, N i Normalizing the aggregate vegetation index for the i-th pel in the aggregate normalized difference vegetation index map.
Step S6: calculating a deviation value of the simulated leaf area index graph relative to the aggregate leaf area index graph, and determining a target sampling area according to the deviation value. The detailed steps are as follows:
(1) Calculating a deviation value of the simulated leaf area index graph relative to the aggregate leaf area index graph by using a preset deviation calculation formula;
wherein, the preset deviation calculation formula is:
in a preset deviation calculation formula, bias is the deviation value, L2 i A simulated leaf area index L1 for an i-th pixel in the simulated leaf area index map i And the aggregate leaf area index of the ith pixel element in the aggregate leaf area index chart is obtained, and n is the total pixel number.
(2) Judging a target sampling area according to preset conditions, wherein the preset conditions are as follows:
where d is a known parameter. d is the desired accuracy, here with reference to the verification accuracy (RMSE: 0.73) of the MODIS leaf area index product, the value of which, d=0.73, is set. And if the deviation value meets a preset condition, determining the corresponding ground areas in the E correlation index graphs as the target sampling areas, otherwise, adding 1 on the basis of the sampling number E to execute step S5.
In the embodiment of the present invention, n=9. If the deviation value meets the preset condition, CI is added 1 、CI 5 、CI 6 And CI (CI) 8 And determining a corresponding ground area in the leaf area index map as a target sampling area. That is, the ground area corresponding to the image shown in fig. 8 is determined as the target sampling area.
Illustratively, CI 1 The leaf area index map corresponds to 9 pixels, namely a pixel where L11 is located, a pixel where L12 is located,... If the ground area corresponding to the pixel where the L11 is located is S1, the ground area corresponding to the pixel where the L12 is located is S2, & gt, the ground area corresponding to the pixel where the L32 is located is S8, and the ground area corresponding to the pixel where the L33 is located is S9. Then CI is 1 Corresponding ground area in leaf area index mapThe combination of this ground area is the target sampling area, S1, S2,..s 8 and S9.
In step S6, if the deviation value does not meet the preset condition, let e=e+1, then in step S4, E (e=4+1=5 in the embodiment of the present invention) correlation indexes are selected from all correlation indexes, and E (e=5) correlation indexes are the smallest correlation indexes among all correlation indexes. In step S5, a new simulated leaf area index map is determined, in step S6, a deviation value of the new simulated leaf area index map with respect to the aggregate leaf area index map is calculated, if the deviation value does not meet a preset condition, e=e+1, and in step S4, E (e=5+1=6 in the embodiment of the present invention) correlation indexes are selected from all correlation indexes, and E (e=6) correlation indexes are the smallest correlation indexes among all correlation indexes. In step S5, a new simulated leaf area index map is determined, in step S6, a deviation value of the new simulated leaf area index map relative to the aggregate leaf area index map is calculated until the deviation value meets a preset condition, and a target sampling area is determined.
It should be noted that the value of the sample number E affects the generation of the computational geometry model and thus the generation of the simulated leaf area index map.
From the above technical solution, the ground sampling method provided by the invention has the following beneficial effects:
(1) The invention provides a leaf area index ground sampling method, which utilizes a correlation index to select a high-representative target sampling area, so that the problems of insufficient sampling, overlapping samples and weak sample representativeness can not occur, and the problem of insufficient samples caused by insufficient prior data can not occur; the correlation indexes are selected from a few to a many, so that the current target sampling area is the minimum sampling area meeting the application requirements, invalid acquisition work cannot occur, and the technical effects of improving the efficiency of the sampling process and meeting the actual application requirements can be achieved;
(2) The leaf area index ground sampling method provided by the invention considers the influence of ground heterogeneity and does not use priori knowledge, and a target sampling area is directly determined on the remote sensing image of the area to be sampled, so that the ground sampling method provided by the invention is simple and efficient, can be directly applied to actually measured collection of leaf area indexes of heterogeneous ground surfaces, and can quickly generate an effective ground sampling scheme under the condition of insufficient prior knowledge; meanwhile, the ground sampling method provided by the invention is applicable to any vegetation type, and is universal, convenient and easy to popularize and use.
While the foregoing is directed to embodiments of the present invention, other and further details of the invention may be had by the present invention, it should be understood that the foregoing description is merely illustrative of the present invention and that no limitations are intended to the scope of the invention, except insofar as modifications, equivalents, improvements or modifications are within the spirit and principles of the invention.
Claims (9)
1. A leaf area index ground sampling method, comprising:
step S1: acquiring a remote sensing image of a region to be sampled, and preprocessing the remote sensing image;
step S2: calculating a normalized difference vegetation index map and a leaf area index map by using the preprocessed remote sensing image;
step S3: carrying out average polymerization treatment on the leaf area index map to obtain a polymerized leaf area index map;
step S4: performing aggregation treatment on the normalized difference vegetation index map to obtain an aggregation normalized difference vegetation index map and an aggregation normalized difference vegetation index variance map; determining a correlation index map by utilizing the aggregate normalized difference vegetation index variance map, and selecting the minimum E correlation indexes;
Step S5: calculating a simulated leaf area index map from the normalized difference vegetation index map, the leaf area index map, the aggregate normalized difference vegetation index and the E correlation indexes;
s5 specifically comprises the following steps:
s5-1: forming a point set to be modeled by target number of normalized difference vegetation indexes and leaf area indexes corresponding to E related indexes, wherein the target number is equal to E multiplied by N, and the obtained result is multiplied by M;
s5-2: calculating an upper convex hull function and a lower convex hull function of the point set to be built by using a preset function, and generating a calculation geometric model by using the upper convex hull function and the lower convex hull function; the computational geometry model is:
L2 i ＝[g ∧ (N i )+g ∨ (N i )]/2
wherein L2 i For the ith pixel value, N in the simulated leaf area index map i G for the ith pixel value, g in the aggregate normalized difference vegetation index map ∧ (x) G is the upper convex hull function ∨ (x) Is the lower convex hull function;
s5-3: sequentially inputting each pixel value in the aggregation normalized difference vegetation index map into the calculation geometric model, and calculating to obtain the simulated leaf area index map;
step S6: calculating a deviation value of the simulated leaf area index graph relative to the aggregate leaf area index graph, and determining a target sampling area according to the deviation value;
S6 specifically comprises the following steps:
s6-1: performing deviation calculation on the simulated leaf area index map and the aggregate leaf area index map to obtain a deviation value;
wherein Bias is the offset value, L2 i L1 is the ith pixel value in the simulated leaf area index map i The ith pixel value in the aggregate leaf area index chart is used, and n is the total pixel number;
s6-2: judging a target sampling area according to preset conditions, wherein the preset conditions are as follows:
and if the deviation value meets a preset condition, determining the ground area corresponding to the E correlation indexes in the leaf area index graph as the target sampling area, otherwise, adding 1 on the basis of the initial sampling number and executing step S5.
2. The leaf area index ground sampling method according to claim 1, wherein S1 specifically comprises:
s1-1: acquiring a remote sensing image of a region to be sampled, wherein the image ground object comprises vegetation type ground objects, and the image wave band comprises a red wave band and a near infrared wave band;
s1-2: performing radiometric calibration on the remote sensing image, and converting the gray value of each pixel in the remote sensing image into a radiation brightness value;
S1-3: performing atmospheric correction on the remote sensing image, and determining the reflectivity value of each pixel in the remote sensing image;
s1-4: and performing geometric correction on the remote sensing image, and determining longitude and latitude information of each pixel in the remote sensing image and the ground area corresponding to each pixel.
3. The leaf area index ground sampling method according to claim 1, wherein S2 specifically comprises:
s2-1: for the remote sensing image, calculating to obtain a normalized difference vegetation index map by utilizing the reflectivity R of a red wave band and the reflectivity NIR of a near infrared wave band;
wherein NDVI represents a normalized difference vegetation index;
s2-2: determining a fitting relation between the leaf area index and the vegetation index;
s2-3: and calculating a leaf area index map by using the fitting relation between the leaf area index and the vegetation index and the normalized difference vegetation index map.
4. A leaf area index ground sampling method according to claim 3, wherein said determining a fitted relationship of leaf area index to vegetation index comprises:
simulating to obtain a plurality of simulated leaf area indexes and simulated vegetation spectrums respectively corresponding to the simulated leaf area indexes by using preset vegetation parameter information and a radiation transmission model;
Calculating a simulated normalized difference vegetation index by using the simulated vegetation spectrum for each simulated vegetation spectrum;
substituting the simulated leaf area index and the corresponding simulated normalized difference vegetation index into a preset formula for each simulated leaf area index, determining a fitting coefficient of the preset formula, and obtaining a fitting relation between the leaf area index and the vegetation index, wherein the preset formula is as follows:
wherein, LAI i For the ith simulated leaf area index, NDVI i Normalizing the differential vegetation index for the ith simulation, and a and b are the fitting coefficients.
5. The leaf area index ground sampling method according to claim 1, wherein the S3 specifically comprises:
s3-1: setting N rows and M columns of aggregation windows, and moving the aggregation windows in the horizontal direction from left to right, wherein the pixels covered on the leaf area index map of the aggregation windows are not overlapped each time; calculating an average value of N multiplied by M leaf area index values in the aggregation window to obtain the aggregation leaf area index;
s3-2: aggregating pixel values of rows 1 to N on the leaf area index map to obtain a first row leaf area index value of the aggregate leaf area index map; and polymerizing pixels of the (n+1) -2N rows on the leaf area index map to obtain leaf area index values of a second row of the polymerized leaf area index map until the last row is polymerized to obtain the polymerized leaf area index map.
6. The leaf area index ground sampling method according to claim 1, wherein S4 specifically comprises:
s4-1: performing aggregation treatment on the normalized difference vegetation index map to obtain an aggregation normalized difference vegetation index map and an aggregation normalized difference vegetation index variance map;
s4-2: determining a related index map by using a fitting relation between the leaf area index and the vegetation index and the aggregate normalized difference vegetation index variance map;
s4-3: the smallest E correlation indexes are chosen.
7. The leaf area index ground sampling method of claim 6, wherein aggregating the normalized difference vegetation index map specifically comprises:
setting N rows and M columns of aggregation windows, and moving the aggregation windows in the horizontal direction from left to right, wherein pixels covered on the normalized difference vegetation index map are not overlapped each time by the aggregation windows; solving an average value of N multiplied by M pixel values in the aggregation window to obtain the aggregation normalized difference vegetation index; solving variances for N multiplied by M pixel values in the aggregation window to obtain the aggregation normalized difference vegetation index variances;
on the normalized difference vegetation index map, the pixel value aggregation processing results of the 1 st to N th rows are pixel values of the aggregation normalized difference vegetation index map and the first row of the aggregation normalized difference vegetation index variance map;
And aggregating pixels of the (n+1) -th row to the (2N) -th row on the normalized difference vegetation index map to obtain pixel values of a second row of the aggregated normalized difference vegetation index map and the aggregated normalized difference vegetation index variance map until the last row is aggregated to obtain the aggregated normalized difference vegetation index map and the aggregated normalized difference vegetation index variance map.
8. The leaf area index ground sampling method of claim 6, wherein the correlation index map determination process specifically comprises:
calculating a correlation index of each pixel by using the aggregate normalized difference vegetation index variance diagram and a correlation index calculation formula to obtain the correlation index diagram, wherein the correlation index calculation formula is as follows:
wherein f (x) =a×e b*x For the fitting relation between the leaf area index and the vegetation index, a and b are fitting coefficients, N i For the vegetation index value of the ith pixel in the aggregate normalized difference vegetation index map, L1 i For the ith pixel value in the aggregate leaf area index map, f' is the second order Taylor expansion formula of f (x), nvar i CI for variance value of ith pixel in the aggregate normalized difference vegetation index variance map i Is the correlation index of the i-th picture element.
9. The method for sampling a leaf area index ground according to claim 6, wherein the selecting process of the E correlation indexes specifically includes:
calculating standard deviation of leaf area indexes corresponding to all pixels in the leaf area index map;
calculating the initial sampling number by using the standard deviation and an initial sampling number calculation formula, wherein the initial sampling number calculation formula is as follows:
and num is the sampling number, the sampling number is equal to the number of related indexes, S is the standard deviation, and Z and d are known parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910231598.8A CN111738974B (en) | 2019-03-25 | 2019-03-25 | Leaf area index ground sampling method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910231598.8A CN111738974B (en) | 2019-03-25 | 2019-03-25 | Leaf area index ground sampling method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111738974A CN111738974A (en) | 2020-10-02 |
| CN111738974B true CN111738974B (en) | 2023-10-03 |
Family
ID=72646081
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910231598.8A Active CN111738974B (en) | 2019-03-25 | 2019-03-25 | Leaf area index ground sampling method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111738974B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103268424A (en) * | 2013-05-31 | 2013-08-28 | 中国科学院遥感与数字地球研究所 | Method for monitoring earth surface drinking water source water gathering area ecology |
| CN103440490A (en) * | 2013-09-16 | 2013-12-11 | 南京大学 | Urban impervious surface extraction method under SAR (Synthetic Aperture Radar) image space-time similarity analysis |
| CN103473613A (en) * | 2013-09-09 | 2013-12-25 | 武汉理工大学 | Landscape structure-surface temperature-electricity consumption coupling model and application thereof |
| CN105303063A (en) * | 2015-12-03 | 2016-02-03 | 武汉大学 | Leaf area index inversion method and system of merged phenological data and remote sensing data |
| CN108520127A (en) * | 2018-03-29 | 2018-09-11 | 华南农业大学 | A kind of EO-1 hyperion inversion method of seeds leaf area index |
| CN108662991A (en) * | 2018-04-08 | 2018-10-16 | 浙江大学 | Plot scale leaves of winter wheat area index evaluation method based on remote sensing satellite data |
-
2019
- 2019-03-25 CN CN201910231598.8A patent/CN111738974B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103268424A (en) * | 2013-05-31 | 2013-08-28 | 中国科学院遥感与数字地球研究所 | Method for monitoring earth surface drinking water source water gathering area ecology |
| CN103473613A (en) * | 2013-09-09 | 2013-12-25 | 武汉理工大学 | Landscape structure-surface temperature-electricity consumption coupling model and application thereof |
| CN103440490A (en) * | 2013-09-16 | 2013-12-11 | 南京大学 | Urban impervious surface extraction method under SAR (Synthetic Aperture Radar) image space-time similarity analysis |
| CN105303063A (en) * | 2015-12-03 | 2016-02-03 | 武汉大学 | Leaf area index inversion method and system of merged phenological data and remote sensing data |
| CN108520127A (en) * | 2018-03-29 | 2018-09-11 | 华南农业大学 | A kind of EO-1 hyperion inversion method of seeds leaf area index |
| CN108662991A (en) * | 2018-04-08 | 2018-10-16 | 浙江大学 | Plot scale leaves of winter wheat area index evaluation method based on remote sensing satellite data |
Non-Patent Citations (1)
| Title |
|---|
| 基于室内光谱反射率的土壤线影响因素分析;刘焕军;张柏;宋开山;王宗明;段洪涛;杨飞;张新乐;;遥感学报(第01期);第1-3页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111738974A (en) | 2020-10-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109884664B (en) | Optical microwave collaborative inversion method and system for urban overground biomass | |
| CN104142142B (en) | Whole world vegetation fraction estimation method | |
| Jawak et al. | Improved land cover mapping using high resolution multiangle 8-band WorldView-2 satellite remote sensing data | |
| CN112381013B (en) | Urban vegetation inversion method and system based on high-resolution remote sensing image | |
| CN110414738B (en) | Crop yield prediction method and system | |
| CN113156439B (en) | SAR wind field and sea wave joint inversion method and system based on data driving | |
| CN115372282B (en) | Farmland soil water content monitoring method based on hyperspectral image of unmanned aerial vehicle | |
| CN114648705A (en) | Carbon sink monitoring system and method based on satellite remote sensing | |
| Abebe et al. | Combined use of Landsat 8 and Sentinel 2A imagery for improved sugarcane yield estimation in Wonji-Shoa, Ethiopia | |
| CN117111092A (en) | High-spatial-resolution remote sensing water quality detection method based on machine learning | |
| CN116343046A (en) | Multi-source remote sensing urban river water quality monitoring method based on self-optimizing machine learning | |
| CN113205014B (en) | Time sequence data farmland extraction method based on image sharpening | |
| CN111738974B (en) | Leaf area index ground sampling method | |
| CN116482317B (en) | Lake water nutrition state real-time monitoring method, system, equipment and medium | |
| CN111879915A (en) | High-resolution monthly soil salinity monitoring method and system for coastal wetland | |
| CN117035066A (en) | Ground surface temperature downscaling method coupling geographic weighting and random forest | |
| CN115082793B (en) | Method and device for rapidly investigating space background condition of forest and grass in water source area | |
| CN113378476B (en) | Global 250-meter resolution space-time continuous leaf area index satellite product generation method | |
| CN115830464A (en) | Plateau mountain agricultural greenhouse automatic extraction method based on multi-source data | |
| CN109460755A (en) | A kind of green light estimation model training method and device, image synthesis method and device | |
| CN117630000B (en) | Soil salt content evaluation method and device, electronic equipment and storage medium | |
| CN116503747B (en) | Heterogeneous earth surface leaf area index inversion method based on multi-scale remote sensing | |
| CN116432849B (en) | Leaf area index prediction method based on primary productivity data assimilation | |
| CN114863053B (en) | Method and device for improving precision of digital elevation model | |
| CN113537604B (en) | Crop yield prediction method and device based on coupling process model and deep learning |
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 |