CN110175931B

CN110175931B - Method for rapidly extracting crop planting area and phenological information in large range

Info

Publication number: CN110175931B
Application number: CN201910388335.8A
Authority: CN
Inventors: 张朝; 骆玉川; 陶福禄; 陈一; 李子悦
Original assignee: Beijing Normal University
Current assignee: Beijing Normal University
Priority date: 2019-05-10
Filing date: 2019-05-10
Publication date: 2020-04-24
Anticipated expiration: 2039-05-10
Also published as: CN110175931A

Abstract

The invention discloses a method for rapidly extracting crop planting area and phenological information in a large range, which comprises the following steps: s1, selecting cultivated land sample grid points, and extracting LAI time sequence data of each sample grid point based on a low spatial resolution remote sensing product; s2: filtering the sample grid point LAI time sequence data to obtain a filtered LAI characteristic curve; s3: extracting a key phenological period of the sample grid points; s4: calculating the Root Mean Square Error (RMSE) of the key phenological period of the sample lattice points extracted by each crop in the research area under different filtering methods and the phenological observation value of the corresponding agricultural gas station, and determining the optimal filtering method according to the RMSE minimum principle; s5: and (4) according to the optimal filtering method, filtering the LAI time sequence data of each cultivated land lattice point in the research area, extracting the key phenological period of the whole growing season of each crop by using the method in the step S3, and identifying the corresponding planting range and area of each crop according to the key phenological period.

Description

Method for rapidly extracting crop planting area and phenological information in large range

Technical Field

The invention relates to the technical field of agricultural remote sensing, in particular to a method for rapidly extracting crop planting area and phenological information in a large range.

Background

The crop planting range, the planting area and the key phenological period are the most basic and important agricultural condition information in the actual agricultural production condition. The phenological period is a key index for reflecting the growth and development stages of crops, and the mastering of the planting distribution condition of the crops is also a necessary condition for carrying out large-scale agricultural monitoring. The currently available crop phenological data mainly come from records of agricultural meteorological stations or laboratory stations, which have authority and authenticity, but the number of sites nationwide is very limited, and the operation and maintenance cost of the agricultural meteorological stations is high, so that the agricultural meteorological stations cannot fully cover a large area. In the actual agricultural production process, farmers often adjust cultivation habits due to factors such as local microclimates and personal living arrangements, so that the phenological information has spatial difference in a small scale range. Therefore, relying solely on records from the agricultural gas station to develop a simulation of a wide range of crop growth process models can be subject to significant uncertainty, often with either over-or under-estimation of the final yield. In contrast, most scholars in recent years try to extract crop phenological information by using remote sensing images, and set a certain rule to identify a crop planting area and invert phenological periods mainly based on remote sensing time sequence data such as NDVI (normalized difference vegetation index) and EVI (incremental virescence index). The remote sensing-based phenological inversion and crop planting area identification method can fully exert the advantage of abundant space-time information of remote sensing data, and has wide application.

However, the current remote sensing-based crop planting area identification and phenological inversion method has the following main disadvantages:

1. the spatial dimensions are small. Currently, most scholars only carry out crop classification and phenological inversion on the basis of high-spatial-resolution remote sensing images in a plot or county scale, and research on province or larger scales is less. Although fine agricultural management can be facilitated, the high-spatial-resolution remote sensing image has the defects of narrow coverage range, short time sequence, large calculation amount and the like, and is difficult to apply in a large-scale range.

2. The extraction method is difficult to apply to a wide area. The existing method for extracting the crop planting area and the phenological climate generally makes rules based on various remote sensing indexes (such as NDVI, EVI, LSWI and the like), and because the extraction methods in different areas have large differences, the rules made under the condition of each area are difficult to be applied to other areas, so that the difficulty and the complexity are increased for the large-scale application.

3. The extraction rule is complex and depends on a large amount of remote sensing data. The existing method for extracting the crop planting area and the phenological climate usually calculates remote sensing indexes based on various remote sensing products, a rule is established according to the size of the mutual relation among the indexes or the threshold range of the remote sensing products, and meanwhile, the remote sensing products often adopt remote sensing images with medium and high spatial resolution, so that the method has the problems that the extraction rule is too complex (a proper threshold needs to be found), the calculated amount is large due to the fact that the remote sensing indexes are too much depended on, and the like.

Therefore, new technologies of general applicability are needed to at least partially address the above limitations in the field of agricultural remote sensing.

Disclosure of Invention

The inventor finds that the advantages of the low-spatial-resolution remote sensing product (such as a GLASS LAI product) and the phenological period record of the agricultural gas station are fully exerted, the advantage of abundant space-time information of the remote sensing product and the authority and accurate advantage of the recorded data of the agricultural gas station are combined, and the large-scale rapid extraction of the crop planting area and the phenological information is realized. The method overcomes the defects that the existing extraction method is only suitable for small scale, the calculated amount is large and the like, and the result extracted by the new method can be well matched with the recorded phenological data of the agricultural gas station and the planting area data of the county-level statistical yearbook.

According to one aspect of the invention, a method for rapidly extracting the phenological information of crops in a large range is provided, which comprises the following steps:

s1, selecting cultivated land sample grid points in the research area according to the land utilization data, and extracting LAI time sequence data of each sample grid point based on a low spatial resolution remote sensing product;

s2: filtering the sample lattice point LAI time sequence data by using a Savitzky-Golay (S-G) filtering method, a double logistic method and a wavelet filter method respectively to obtain a filtered LAI characteristic curve;

s3: extracting a sample grid point key phenological period: according to the phenological observation value recorded by the agricultural gas station, firstly defining the phenological extraction time range of the phenological period of crops adjacent to the sample lattice point of the agricultural gas station, and then adopting a phenological extraction algorithm to identify the inflection point and the peak value of an LAI characteristic curve as the key phenological period of the sample lattice point extracted by remote sensing;

s4: calculating the Root Mean Square Error (RMSE) of the key phenological period of the sample lattice points extracted by each crop in the research area under different filtering methods and the phenological observation value of the corresponding agricultural gas station, and determining the optimal filtering method according to the RMSE minimum principle;

s5: and (4) according to the optimal filtering method, filtering the time sequence data of each cultivated land lattice point LAI in the research area, and extracting the key phenological period of the whole growing season of each crop by using the method of the step S3.

According to an embodiment of the invention, the low spatial resolution remote sensing product may be derived from the GLASS LAI remote sensing product (1km × 1km), or may be other low spatial resolution products such as the NOAA/AVHRR remote sensing product; the land use data may be, for example, from products such as land use/cover change data (LUCC) and national land use data (NLCD), or other suitable sources.

According to the embodiment of the present invention, in step S1, Savitzky-Golay (S-G) is used as a low-pass filtering method for smoothing time series data by using a local polynomial regression model, and the formula is shown in (1):

wherein Y is the original LAI value, Y^*Is the filtered LAI value, j refers to the jth point of the LAI timing sequence, C_iIs the filter coefficient of the ith data point, m is the window radius, and N is the width of the sliding window, i.e., 2m + 1.

According to the embodiment of the present invention, in step S2, the Double logistic used is a method of combining local fitting to obtain an overall fitting result, and in the local fitting process, the function used is a Double logistic function, and the formula is shown in (2):

in the formula, x₁Is the position of the left inflection point, x₂Is the rate of change of the curve at the left-hand corner, x₃Position of right inflection point, x₄Is the curve rate of change at the right turn point; t is the index of the LAI time series.

According to the embodiment of the present invention, in step S2, the wavelet filtering is an effective method for reducing the noise of time series information, and the wavelet function, the wavelet transform and the signal reconstruction formula are shown in (3) - (5):

in the formula (I), the compound is shown in the specification,

is a wavelet function.

Wherein f (t) is the input signal, a is the scale parameter, b is the displacement parameter,

the wavelet mother functions (Daubechies, Coiflet and Symlet are used, and these wavelet mother functions are well known in the art and are not described herein in detail).

Wherein W is the sum of a series of wavelets of different widths, W (a, b)_iFor the ith wavelet transform constructed for equation (4), x is the number of wavelet transforms.

According to an embodiment of the present invention, in step S2, the sample grid point is selected from grid points where the two land use data are collectively identified as arable land.

According to an embodiment of the present invention, in step S3, the phenological extraction algorithm extracts phenological information according to equations (6) to (8):

in the formula, LAI' is a first derivative of the LAI time series, t1 is a first inflection point identified, and Period is a corresponding phenological extraction time range defined according to phenological periods recorded by grid points adjacent to the agricultural gas station.

LAI(tmaX)＝max{LAI_i|i∈Period} (7)

Wherein, { LAI_iI belongs to Period and is recorded in the neighborhood of the agricultural gas station according to the lattice pointThe corresponding object-to-object defined by the candidate extracts the LAI time series within the time range, tmax being the identified peak.

Where { LAI' (i) | i ∈ Period } is the LAI first derivative time series within the corresponding phenology extraction time range defined from the phenology Period recorded by the grid point neighboring the agricultural gas station, and t2 is the second inflection point identified.

According to the embodiment of the invention, in step S4, after removing the grid points which are identified as non-cultivated lands by two pieces of land utilization data together by using mask processing, the remaining grid points are used as cultivated land grid points of the to-be-extracted phenological information.

According to the embodiment of the present invention, the key phenological period may be selected from a green-turning period, a tillering period, a trefoil period, a transplanting period, a heading period and a mature period according to the kind of the crop, and in step S5, the lattice points that completely meet the algorithm extraction rules and can simultaneously extract three key phenological periods of the whole growing season of a certain crop are used as the crop planting lattice points of the crop, wherein when the crop is winter wheat, only two key phenological periods of the green-turning period and the heading period may be required to be simultaneously extracted.

According to embodiments of the invention, the crop may be winter wheat, spring wheat, summer corn, spring corn, one season rice, early rice and late rice.

According to an embodiment of the invention, the region of interest is a region of above provincial level. For example, it may be a city or provincial administrative division, or a plurality of provinces, cities, or even nationwide or worldwide.

According to another aspect of the present invention, there is provided a method for rapidly extracting the planting area of crops in a wide range, which comprises using the key phenological period obtained by the above method according to the present invention to further identify the corresponding planting area of each crop.

Compared with the existing method for extracting the crop planting area and the phenological climate, the method overcomes the defects that the traditional method is only suitable for small scale, large in calculated amount and the like; according to the invention, based on the low spatial resolution remote sensing data and the climate data recorded by the agricultural gas station, the crop planting range, area and climate information can be rapidly extracted in a large-range area, and the reliability of the result can be ensured.

In addition, the method has high repeatability, and is favorable for realizing large-scale popularization and application of business. The embodiment case performs example operation on the whole China by using low-spatial-resolution remote sensing products and the recorded data of the gas station, and provides a model for extracting the climate, the planting range and the planting area of the global crops.

Drawings

The same reference numbers in the drawings identify the same or similar elements or components. The objects and features of the present invention will become more apparent in view of the following description taken in conjunction with the accompanying drawings, in which:

fig. 1 is a schematic flow chart of a method for rapidly extracting crop planting area and climate information in a large range based on a low spatial resolution remote sensing product according to an embodiment of the invention.

Fig. 2 is a schematic diagram of the phenological extraction precision result of the method for rapidly extracting the crop planting area and the phenological information in a large range based on the low spatial resolution remote sensing product according to one embodiment of the invention.

Fig. 3 is a schematic diagram of the result of the planting area extraction accuracy of the method for rapidly extracting the planting area and the phenological information of the crops in a large range according to one embodiment of the present invention.

Fig. 4 is a spatial distribution diagram of the phenological extraction result of the method for rapidly extracting the crop planting area and phenological information in a wide range according to one embodiment of the present invention.

Fig. 5 is a spatial distribution diagram of a planting range of a method for rapidly extracting a crop planting area and phenological information over a wide range according to an embodiment of the present invention.

Detailed Description

For a clear description of the solution according to the invention, preferred embodiments are given below and are described in detail with reference to the accompanying drawings. The following description is merely exemplary in nature and is not intended to limit the present disclosure, application, or uses

Fig. 1 is a schematic flow chart of a method for rapidly extracting crop planting area and climate information in a large range based on a low spatial resolution remote sensing product according to an embodiment of the invention. The process of the invention is further illustrated below with reference to a specific example, with reference to fig. 1:

the whole China is taken as a research area, the research objects are spring wheat, winter wheat, spring corn, summer corn, one-season rice, early rice and late rice, and the research time is 2015 year with 2000-plus materials. Processing a 1km × 1km GLASS LAI remote sensing product into LAI time sequence data of grid points by grid points year after year, selecting sample grid points from grid points which are identified as cultivated land by the two land utilization products based on a 1km × 1km LUCC time-space data set (published by China academy of sciences) and an NLCD (land utilization/land cover data set from a China resource environment database), unifying the coordinate projection of all data into Asia North Albers Equal Area Conic, and taking the spatial reference as WGS 1984.

The selection of the sample grid points can be performed by referring to the distribution and the number of the specific agricultural gas stations, so that the selection of the subsequent optimal filtering method can be facilitated. For the sample grid points, the LAI time series data were filtered using Savitzky-Golay (S-G) filtering (filter window size set to 3, 4, 5, respectively), Double logistic, and wavelet filter (wavelet mother functions Daubechies, Coiflet, and Symlet, respectively) methods, respectively, on a grid-by-grid basis.

The Savitzky-Golay (S-G) is a low-pass filtering method for smoothing time series data by using a local polynomial regression model, and formula (1) is:

wherein, the original LY is the original LAI value, Y^*Is the filtered LAI value, j refers to the jth point of the LAI timing sequence, C_iIs the filter coefficient of the ith data point, m is the window radius, and N is the width of the sliding window, i.e. 2m +1

The adopted Double logistic is a method for combining local fitting to obtain an overall fitting result, in the process of local fitting, the adopted function is a Double logistic function, and the formula (2) is as follows:

in the formula, x₁Is the position of the left inflection point, x₂Is the rate of change of the curve at the left-hand corner, x₃Position of right inflection point, x₄For the rate of change of the curve at the right-hand corner, t is an index of the LAI time series, and may be, for example, 1,2, …,46, etc. according to the amount of data to be fitted.

The adopted wavelet filtering is an effective method for reducing the time series information noise, and the wavelet function, the wavelet transformation and the signal reconstruction formula are as follows (3) to (5):

in the formula (I), the compound is shown in the specification,

is a wavelet function.

is a wavelet mother function.

Wherein W is the sum of a series of wavelets of different widths, W (a, b)_iThe ith wavelet transform constructed for equation (4), x being the number of wavelet transforms,

where t is the index of the LAI time series.

Different planting systems, such as single-season rice, spring corn, and double-season early and late rice, winter wheat and summer corn, etc., have different LAI curve characteristics, i.e., the key phenological stages may differ from crop to crop. For example, specifically: obvious inflection points (namely nodes of LAI values which are changed from low level to continuous rising stage for a long time) can be observed in the wheat turning stage or the tillering stage, the corn stage and the rice transplanting stage; all crops reach the peak value of an LAI curve in the heading stage; a second inflection point (i.e., the node where the rate of reduction of LAI is greatest) is observed in all crops during the mature period. The same crop may have multiple key phenological stages, for example, rice may include a transplanting stage, heading stage, and maturity stage.

Then, determining the possible range of the crop phenological period according to the phenological observation value of the agricultural gas station adjacent to the sample grid point, for example, taking the time length of three-period remote sensing data which is 24 days before and after the average value of the current year or multiple years of observation values as the phenological time extraction range, that is, determining the phenological time extraction range of each key phenological period; and identifying inflection points and peak values of the LAI time sequence curve by adopting a phenological extraction algorithm from year to year and lattice points, wherein a specific formula (6) is as follows:

in formula (6), LAI' is a first derivative of the LAI time series, t1 is a first inflection point identified, such as DOY1 in fig. 1, and Period is a phenological time extraction range defined according to a corresponding phenological Period recorded in the vicinity of the agricultural gas station at the grid point, such as the phenological time extraction range which may be the first inflection point (e.g., rice transplanting Period) in fig. 1.

LAI(tmaX)＝max{LAI_i|i∈Period} (7)

In the formula (7), { LAI_iI e Period is the LAI time series in the phenological extraction time range defined according to the corresponding phenological Period recorded by the grid point adjacent to the agricultural gas station, tmax is the identified peak value, such as DOY2 in fig. 1, in which case Period may be the phenological extraction time range of the peak value (such as rice heading Period) in fig. 1.

In equation (8), { LAI' (i) | i ∈ Period } is a LAI first derivative time series in a phenological extraction time range defined according to the phenological Period recorded by the grid point adjacent to the agricultural gas station, and t2 is a second inflection point identified, for example, DOY3 in fig. 1, where Period may be, for example, the phenological time extraction range of the second inflection point (e.g., rice maturity) in fig. 1.

When the crop is a dual season crop, and the entire growing season thereof includes two seasons, the extracted phenological information may include two first inflection points DOY1, two second inflection points DOY3, and two peak DOY 2.

Calculating the key phenological period of sample lattice points extracted by various provinces and crops in the research area under different filtering methods and the Root Mean Square Error (RMSE) of observation values of corresponding agricultural gas stations, and determining the optimal filtering method according to the RMSE minimum principle; it will be appreciated that different crops and different regions may have different optimal filtering methods, and that the results calculated for each study area according to the above embodiments of the invention are shown in table 1.

TABLE 1 optimal filtering method for different crops in different areas

Winter wheat

Double cropping rice

One season rice

Summer corn

Spring corn

Spring wheat

Anhui badge

SG-3

SG-5

SG-4

Gansu (Gansu)

SG-3

SG-5

SG-3

SG-5

Hainan province

db8

Heilongjiang

SG-5

Henan province

SG-3

SG-5

SG-4

Jiangsu

SG-3

SG-5

SG-4

db4

Jilin

SG-5

SG-4

Inner Mongolia

SG-5

SG-3

Qinghai (Qinghai-food)

db3

SG-5

DL

Shandong (mountain east)

SG-3

SG-5

SG-4

Shaanxi province

SG-3

SG-5

SG-3

Shanxi province

SG-3

SG-5

SG-3

SG-5

DL

Yunnan province

SG-3

SG-5

SG-3

Zhejiang river

SG-4

SG-3

SG-3, SG-4 and SG-5 all represent S-G filtering methods, and the difference is that the sizes of filtering windows are respectively set to be 3, 4 and 5; DL represents Double logistic filtering method; db3, db4, db8 each represent a wavelet filter having a wavelet mother function Daubechies, with the difference that the vanishing moment order is 3, 4, 8, respectively. The vanishing moment order is defined as formula (9):

in the formula (9), the reaction mixture is,

is a waveletThe number is p is more than or equal to 0 and less than N. The wavelet function is said to have an vanishing moment of order N,

orthogonal to any polynomial of order n-1.

The open white in table 1 indicates that the crop was not planted in the area.

After the optimal filtering method is determined, mask processing is utilized to remove grid points of the two land utilization data which are identified as non-cultivated lands together, the remaining grid points are used as cultivated land grid points of to-be-extracted phenological information, filtering processing is carried out on LAI time sequence data of the cultivated land grid points according to the optimal filtering method of each crop in each province, and then the key phenological period of the whole growing season of the crop is extracted by the method of the step S3, wherein the grid points which completely accord with the algorithm extraction rule and can simultaneously extract three key phenological periods of the whole growing season are used as crop planting grid points. When the crop is winter wheat, two key phenological periods of a green turning period and a heading period can be extracted at the same time.

After determining the key phenological period for each crop, the crop planting area (planting area) can be further identified accordingly.

The method for rapidly extracting the crop planting area and the phenological information in the large range is implemented, combines the abundant advantages of the spatial-temporal information of the remote sensing data and the authoritative and accurate advantages of the recorded data of the agricultural gas station, achieves rapid extraction of the crop planting area and the phenological information in the large range, and overcomes the defects that the existing extraction method is only suitable for small scale, large in calculated amount and the like. The results extracted by the new method can be well matched with the recorded phenological data of the agricultural gas station and the planting area data of the county-level statistical yearbook (see table 2, table 3, and fig. 2 and 3 specifically). The error of the phenological period recognized by each crop is mostly concentrated within 10 days (73.0%), the average RMSE is less than 10 days, and the coefficient R is determined²About 0.97, considering that the error of the agricultural gas station record and the step length of the remote sensing data can influence the object recognition result, the error isIs completely acceptable. The RMSE extracted from the planting area of each crop is respectively 9.68kha, 9.86kha and 7.04kha, the relative root mean square difference RRMSE is within 40 percent, and the coefficient R is determined²Are all larger than 0.79, and the error is completely acceptable in consideration of the error of the statistic yearbook record

TABLE 2 extraction accuracy results for various crop key phenological periods

Crops	RMSE_DOY1	RMSE_DOY2	RMSE_DOY3	RMSE_mean	R²
						Rice (Oryza sativa L.) with improved resistance to stress	6.53	8.36	9.02	7.97	0.98
Wheat (Triticum aestivum L.)	5.37	8.55	11.46	8.46	0.97
						Corn (corn)	5.24	9.48	10.20	8.31	0.97

TABLE 3 result of extraction accuracy of various crop planting areas

Crops	RMSE(Kha)	RRMSE(％)	R²
				Rice (Oryza sativa L.) with improved resistance to stress	9.68	37.03	0.80
Wheat (Triticum aestivum L.)	9.86	37.36	0.79
				Corn (corn)	7.04	35.66	0.85

The method is based on the low-spatial-resolution remote sensing product and the climate record of the agricultural gas station, can quickly extract the crop planting area and the climate information in a large spatial scale range, has high extraction efficiency and accurate extraction result, has universality, and overcomes the defects that the traditional method is only suitable for small scale, has large calculated amount and the like.

The principles and embodiments of the present invention have been described herein using specific examples, which are presented solely to aid in the understanding of the apparatus and its core concepts; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.

Claims

1. A method for rapidly extracting phenological information of crops in a large range comprises the following steps:

s1: selecting cultivated land sample grid points in a research area according to land utilization data, and extracting LAI time sequence data of each sample grid point based on a low spatial resolution remote sensing product;

s2: filtering the sample lattice point LAI time sequence data by using a Savitzky-Golay filtering method, a Double logistic method and a wavelet filter method respectively to obtain a filtered LAI characteristic curve;

s4: calculating the root mean square error RMSE of the sample lattice point key phenological period extracted by each crop in the research area under different filtering methods and the phenological observation value of the corresponding agricultural gas station, and determining the optimal filtering method according to the RMSE minimum principle;

s5: according to the optimal filtering method, filtering time sequence data of each cultivated land lattice point LAI in the research area, and extracting key phenological periods of the whole growing season of each crop by using the method of the step S3;

in step S3, the phenological extraction algorithm extracts phenological information according to the formulas (6) to (8):

in the formula, LAI' is a first derivative of the LAI time sequence, t1 is a first identified inflection point, and Period is a corresponding phenological extraction time range defined according to phenological periods recorded by grid points adjacent to the agricultural gas station;

LAI(tmax)＝max{LAI_i|i∈Period} (7)

wherein, { LAI_iI belongs to Period and is an LAI time sequence in a corresponding phenological extraction time range defined according to phenological periods recorded by grid points adjacent to the agricultural gas station, and tmax is an identified peak value;

2. The method for the large-scale rapid extraction of the phenological information of crops as claimed in claim 1, wherein the low spatial resolution remote sensing product is GLASS LAI, 1km x 1 km.

3. The method for extracting phenological information of crops in a large scale in a rapid manner according to claim 1, wherein in step S2, Savitzky-Golay is a low-pass filtering method for smoothing time series data by using a local polynomial regression model, as shown in formula (1):

wherein Y is the original LAI value, Y^*For filtered LAI valuesJ denotes the j point of the LAI timing sequence, C_iIs the filter coefficient of the ith data point, m is the window radius, and N is the width of the sliding window, i.e. 2m + 1;

double localization is a method for combining local fitting to obtain an overall fitting result, and in the local fitting process, the adopted function is a Double localization function, as shown in formula (2):

in the formula, x₁Is the position of the left inflection point, x₂Is the rate of change of the curve at the left-hand corner, x₃Position of right inflection point, x₄Is the curve rate of change at the right turn point; t is LAI timing index;

wavelet filtering is a method for reducing noise of time series information, in which a wavelet function, wavelet transformation, and signal reconstruction are shown in equations (3) to (5):

in the formula (I), the compound is shown in the specification,

is a wavelet function;

selecting Daubechies, Coiflets and Symlet as wavelet mother functions; t is LAI timing index;

wherein W is a seriesColumn sum of wavelets of different widths, W (a, b)_iFor the ith wavelet transform constructed for equation (4), x is the number of wavelet transforms.

4. The method for rapid wide-scale extraction of phenological information of agricultural crops as claimed in claim 1, wherein in step S1, the land use data are derived from both LUCC and NLCD products, and the sample lattice point is selected from lattice points identified together as cultivated land by said two land use data.

5. The method for extracting the phenological information of the crops in a large scale according to claim 4, wherein the step S5 includes removing the grid points which are identified as the non-cultivated land together by the two land utilization data by masking treatment, and then using the rest grid points as the cultivated land grid points of the phenological information to be extracted.

6. The method of claim 1, wherein the key phenological period is selected from a green-turning period, a tillering period, a trefoil period, a transplanting period, a heading period and a mature period, and in step S5, three cultivated land lattice points that completely meet the algorithm extraction rules and can simultaneously extract three key phenological periods of a whole growing season of a certain crop are used as the crop planting lattice points, wherein if the crop is winter wheat, the green-turning period and heading period can be simultaneously extracted.

7. The method for the wide-scale rapid extraction of crop phenological information according to claim 1, wherein the crop is selected from winter wheat, spring wheat, summer corn, spring corn, first season rice, early season rice and late season rice.

8. The method for extracting phenological information of crops in a wide range in claim 1, wherein the research area is an area range above provincial level.

9. A method for rapidly extracting the planting area of crops in a large range, which comprises identifying the corresponding planting area of each crop by using the key phenological period obtained by the method according to any one of claims 1 to 8.