CN117760984A - Winter wheat SPAD space-time change monitoring method - Google Patents
Winter wheat SPAD space-time change monitoring method Download PDFInfo
- Publication number
- CN117760984A CN117760984A CN202311794390.XA CN202311794390A CN117760984A CN 117760984 A CN117760984 A CN 117760984A CN 202311794390 A CN202311794390 A CN 202311794390A CN 117760984 A CN117760984 A CN 117760984A
- Authority
- CN
- China
- Prior art keywords
- spad
- winter wheat
- texture
- feature
- bands
- 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.)
- Pending
Links
- 241000209140 Triticum Species 0.000 title claims abstract description 104
- 235000021307 Triticum Nutrition 0.000 title claims abstract description 102
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000012544 monitoring process Methods 0.000 title claims abstract description 26
- 230000008859 change Effects 0.000 title abstract description 12
- 230000003595 spectral effect Effects 0.000 claims abstract description 55
- 230000012010 growth Effects 0.000 claims abstract description 43
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 32
- 230000004927 fusion Effects 0.000 claims abstract description 15
- 238000002310 reflectometry Methods 0.000 claims abstract description 14
- 239000011159 matrix material Substances 0.000 claims abstract description 3
- 229930002875 chlorophyll Natural products 0.000 claims description 20
- 235000019804 chlorophyll Nutrition 0.000 claims description 20
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 claims description 20
- 238000001228 spectrum Methods 0.000 claims description 13
- 230000005855 radiation Effects 0.000 claims description 12
- 238000012795 verification Methods 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 238000010801 machine learning Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 238000012706 support-vector machine Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000017260 vegetative to reproductive phase transition of meristem Effects 0.000 abstract description 17
- 230000001850 reproductive effect Effects 0.000 abstract description 7
- 239000000284 extract Substances 0.000 abstract description 2
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 22
- 229910052757 nitrogen Inorganic materials 0.000 description 11
- 238000010187 selection method Methods 0.000 description 10
- 238000012216 screening Methods 0.000 description 9
- 238000005259 measurement Methods 0.000 description 7
- 238000005070 sampling Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 238000010200 validation analysis Methods 0.000 description 4
- 230000003698 anagen phase Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 230000008030 elimination Effects 0.000 description 3
- 238000003379 elimination reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000014759 maintenance of location Effects 0.000 description 3
- 235000016709 nutrition Nutrition 0.000 description 3
- 230000035764 nutrition Effects 0.000 description 3
- 230000000717 retained effect Effects 0.000 description 3
- 238000009331 sowing Methods 0.000 description 3
- 238000011282 treatment Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000009395 breeding Methods 0.000 description 2
- 230000001488 breeding effect Effects 0.000 description 2
- 235000013339 cereals Nutrition 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 239000003337 fertilizer Substances 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 239000000618 nitrogen fertilizer Substances 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 241000196324 Embryophyta Species 0.000 description 1
- 102100033398 Glutamate-cysteine ligase regulatory subunit Human genes 0.000 description 1
- 101000870644 Homo sapiens Glutamate-cysteine ligase regulatory subunit Proteins 0.000 description 1
- 241000208125 Nicotiana Species 0.000 description 1
- KWYUFKZDYYNOTN-UHFFFAOYSA-M Potassium hydroxide Chemical compound [OH-].[K+] KWYUFKZDYYNOTN-UHFFFAOYSA-M 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000035558 fertility Effects 0.000 description 1
- 230000004720 fertilization Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002686 phosphate fertilizer Substances 0.000 description 1
- 239000004069 plant analysis Substances 0.000 description 1
- 229940072033 potash Drugs 0.000 description 1
- 235000015320 potassium carbonate Nutrition 0.000 description 1
- BWHMMNNQKKPAPP-UHFFFAOYSA-L potassium carbonate Substances [K+].[K+].[O-]C([O-])=O BWHMMNNQKKPAPP-UHFFFAOYSA-L 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a winter wheat SPAD space-time change monitoring method, and relates to the technical field of crop monitoring. The invention comprises the following steps: the unmanned aerial vehicle multispectral sensor acquires winter wheat canopy images in the heading period, the flowering period and the later grouting period, extracts spectral features based on the reflectivity of the winter wheat canopy and texture features based on a gray level co-occurrence matrix from the multispectral images, adopts a feature selection strategy to select sensitive remote sensing features, and applies a feature fusion strategy and an SVR algorithm to construct a winter wheat SPAD estimation model. The spectral characteristics of the combination of the near infrared band and other bands can fully capture the spectral difference of the winter wheat SPAD in the reproductive growth stage, and the texture characteristics of the red band and the near infrared band are more sensitive to the winter wheat SPAD. The stability and the estimation precision of the SPAD model constructed by applying the feature selection strategy and the feature fusion strategy are superior to those of models which independently apply the feature strategy and do not apply the strategy.
Description
Technical Field
The invention belongs to the technical field of crop monitoring, and particularly relates to a winter wheat SPAD space-time change monitoring method.
Background
Soil Plant Analysis Development (SPAD) characterizes the relative content of chlorophyll in leaves, which is an important index reflecting the growth and development of crops and is closely related to the nutrition status of crops. Therefore, the accurate and efficient acquisition of the winter wheat SPAD has important significance for field management decision and growth condition monitoring.
Traditional methods for obtaining chlorophyll content of crops mainly rely on field destructive sampling and indoor chemical analysis, and are time-consuming, labor-consuming and high in cost. Studies have demonstrated that the data obtained with a handheld SPAD-502 chlorophyll meter is closely related to chlorophyll content by chemical analysis in the laboratory. Compared with the traditional measurement method of chlorophyll, the SPAD acquisition mode is more convenient, can carry out nondestructive sampling, but because of the limitation of the measurement point, the measurement work still has certain labor cost, the operation scale is small, the efficiency is low, and the requirement of acquiring the chlorophyll content information of crops in a large area is difficult to meet. SPAD is an important index for evaluating crop nutrition status, and is an important parameter for representing reproductive growth state from winter wheat heading stage to grouting stage. Therefore, the nondestructive, rapid and accurate monitoring of the winter wheat SPAD has an important effect on guaranteeing stable grain yield and guiding the accurate management in the field.
Spectral features (vegetation index, band reflectivity, etc.) calculated from unmanned aerial vehicle images have become common remote sensing information in the field of precision agriculture, and proved to have great potential in estimating phenotypic traits of crops LAI, AGB, SPAD, nitrogen content, etc. However, in the case of high canopy coverage in the late stage of crop growth, the vegetation index may produce spectral saturation effects. This is due to the complex background of the wheat canopy during the reproductive growth phase, the unmanned aerial vehicle image contains ears, stems, leaves and small amounts of soil, and saturated image information reduces the sensitivity of the vegetation index. Thus, it is difficult to build a reliable wheat SPAD estimation model using spectral features alone during the reproductive growth phase.
In the prior art, the estimation of SPAD is mainly focused on the nutrition growth stage of wheat, and the estimation model has low accuracy because winter wheat has a complex canopy background in the later growth stage. In view of the current state of the art, there has been a great deal of research involving fusion applications of spectral features and texture features, but for this vast fused dataset, the choice of feature variables is rarely considered, and the redundancy of data between spectral features and texture features can degrade model performance.
Disclosure of Invention
The invention aims to provide a method for monitoring space-time change of winter wheat SPAD, which takes winter wheat with different varieties and different nitrogen application amounts as a research object, acquires winter wheat canopy images in the heading period, the flowering period and the post-grouting period by using an unmanned aerial vehicle multispectral sensor, extracts spectral features and texture features of the images by adopting an image analysis technology, screens remote sensing indexes sensitive to winter wheat SPAD respectively based on 2 feature selection methods of Borata and Recursive Feature Elimination (RFE), and constructs a SPAD prediction model of winter wheat in the reproductive growth stage by using a Support Vector Regression (SVR) machine learning algorithm. Coefficient of determination (R 2 ),root mean square error(RMSE)and residual prediction deviationThe (RPD) evaluates the 63 models, solving the existing problem.
In order to solve the technical problems, the invention is realized by the following technical scheme:
a winter wheat SPAD space-time change monitoring method comprises the following steps:
step S001: collecting multispectral data of winter wheat in a growing period by using an unmanned aerial vehicle;
step S002: determining the relative chlorophyll content value (SPAD) of winter wheat in a plurality of plots by using SPAD-502 type chlorophyll meter;
step S003: unmanned aerial vehicle image preprocessing: the multispectral images of the unmanned aerial vehicle collected in each growth period are imported into PIX4 Dapp er software for image stitching;
step S004: calculating the average number of the relative chlorophyll content values of each land block, analyzing the average number, and respectively constructing a SPAD prediction model by using the spectrum characteristics, the texture characteristics and the spectrum texture characteristic fusion set;
step S005: the chlorophyll content of winter wheat based on SPAD value is predicted by adopting a support vector machine regression (support vector machine regression, SVR) algorithm based on R language (version 4.1.3).
Further, the unmanned aerial vehicle includes 5 1208 ten thousand pixel monochrome sensors, and monochrome sensors are respectively: the central wavelengths of the monochromatic sensors are 450nm,560nm,650nm,730nm and 840nm respectively, the bandwidths of the monochromatic sensors are +/-16 nm, +/-16 nm and +/-26 nm respectively, the unmanned aerial vehicle comprises an RTK (real-time kinematic) system, and the vertical precision of the RTK (real-time kinematic) system is +/-1.5 cm and the horizontal precision of the RTK (real-time kinematic) system is +/-1 cm.
Further, in the step S003, the image stitching method includes the following steps:
step S0311: aligning the histogram by utilizing a characteristic point matching algorithm, and acquiring a dense point cloud and a texture grid based on the unmanned aerial vehicle image and the position data;
step S0312: correcting the difference between the wave bands;
step S0313: creating a vector (shape) file by using ArcGIS10.2, dividing a sample area according to an orthophoto map of a test area, generating a plurality of vector cells overlapped on a winter wheat image, and giving ID information to the vector cells one by one.
Further, the step S003 further includes the steps of:
step S031: spectral feature extraction: acquiring a plurality of original wave band reflectivities of a winter wheat growing period canopy from a multispectral image;
step S032: texture feature extraction: and extracting texture information of blue, green, red edges and near infrared bands in the winter wheat canopy multispectral image by using a gray level co-occurrence matrices (GLCM).
Further, in the step S0312, the method for correcting the difference between the bands is as follows:
a plurality of radiation calibration plates with known reflectivity are distributed and controlled on the ground;
the multispectral image is radiation corrected using ground radiation reticle image information of known reflectivity based on an empirical linear method (empirical line method).
Further, in the step S031, 30 spectral characteristics for crop growth monitoring and parameter evaluation are selected, and the spectral characteristics are classified into 5 categories, which are respectively 5 original bands, 7 vegetation indexes consisting of only visible light bands, 5 vegetation indexes consisting of red-edge bands and no near-infrared band, 8 vegetation indexes consisting of near-infrared bands and no red-edge band, and 5 vegetation indexes consisting of both near-infrared and red-edge bands.
Further, in the step S032, the texture information includes eight indexes including a Mean (Mean), a Variance (Variance), a Homogeneity (Homogeneity), a Contrast (Contrast), a Dissimilarity (Dissimilarity), an Entropy (Entropy), an angular Second moment (Second moment) and a Correlation (Correlation), and a maximum value, a minimum value, a Mean value and a standard deviation of each texture index (GLCM) are calculated to obtain 5×8×4 texture indexes.
Further, in the step S004, when the SPAD prediction model is constructed, different feature parameter sets are screened by adopting an RFE algorithm and a Boruta algorithm, and the optimal feature combination is obtained by fusing the different feature parameter sets and comparing and analyzing the model performances of the feature parameter sets which are not screened;
the feature parameter set variable screened by the RFE algorithm takes R-as prefix, and the feature parameter set screened by the Borata algorithm is respectively expressed by C-and CT-;
the feature parameter set screened by the Borata algorithm comprises two types of "Confirmed" (C) and "pending" (CT).
Further, in the step S005, the method for predicting the chlorophyll content of the winter wheat based on the SPAD value comprises the following steps:
step S0051: dividing the SPAD characteristic parameter set of winter wheat in each growth period into a training set and a verification set;
step S0052: 3 statistical indexes are adopted to test a machine learning model, and the 3 statistical indexes are respectively: determining coefficient (R) 2 ) Root Mean Square Error (RMSE) and performance to bias Ratio (RPD); wherein:
in the method, in the process of the invention,and y i For the observations and measured values of SPAD, +.>The average value of the SPAD observations is represented by n, the number of samples, and SD, the standard deviation of the observations.
The invention has the following beneficial effects:
according to the invention, the spectral difference of the winter wheat SPAD in the reproduction growth stage can be fully captured through the spectral features of the near infrared band participation combination, and the texture features of the red light band and the near infrared band extracted by the Grey Level Co-occurrence Matrices (GCLM) method are sensitive to the wheat SPAD. In addition, the robustness and the accuracy of the prediction model are remarkably improved by combining the feature selection strategy with the feature fusion strategy.
Of course, it is not necessary for any one product to practice the invention to achieve all of the advantages set forth above at the same time.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for monitoring space-time variation of winter wheat SPAD;
FIG. 2 is a graph showing the spectral characteristics and space-time variation of SPAD of the present invention for three growth phases of winter wheat;
FIG. 3 is a graph of wheat spectral characteristics versus SPAD fold line change at 4 different nitrogen application levels;
FIG. 4 is a graph of the spatial-temporal variation of texture features versus SPAD for three different nitrogen treatments during the reproductive period;
FIG. 5 is a graph of wheat texture characteristics versus SPAD fold line change at 4 nitrogen application levels;
FIG. 6 is a graph of the optimal variable numbers selected by the Borata and RFE feature selection methods for different feature sets during different breeding periods;
FIG. 7 is a scatter plot of an optimal model for estimating SPAD in the late growth stage of winter wheat.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
It should be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It should also be understood that the term "and/or" as used in this specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.
Reference in the specification to "one embodiment" or "some embodiments" or the like means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," and the like in the specification are not necessarily all referring to the same embodiment, but mean "one or more but not all embodiments" unless expressly specified otherwise. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless expressly specified otherwise.
Embodiment one:
the invention relates to a method for monitoring space-time change of winter wheat SPAD, which is developed in Chuzhou city of Anhui province in China in 2020-2021 (32 DEG 48 '52' N,117 DEG 46 '7' E), is positioned at the middle and downstream of Yangtze river, belongs to subtropical monsoon climate, is moist in climate, is clear in four seasons, has an average annual temperature of 15.4 ℃, has an annual precipitation of 1000-1100mm, has an average annual precipitation daily number of 144d, and has an annual frost-free period of about 210 d.
This time in the fieldThe test has 36 plots, each of 16m 2 (2X 8 m) involving 3 winter wheat varieties with stable high yield potential (V1: huai wheat 44, V2: nicotiana 999 and V3: ning Mai 13) and 4 nitrogen fertilizer application levels (N0-N3: 0, 100, 200, 300 kg/ha), each treatment was repeated 3 times. The fertilization scheme is to apply phosphate fertilizer (P=90 kg/ha) and potash fertilizer (K=135 kg/ha) as base fertilizers before sowing, and the nitrogen fertilizer application ratio before sowing and in the jointing period is 6:4. the sowing date is 11/7/2020, the winter wheat is manually planted in a drill mode, the row spacing is 30cm, the winter wheat is harvested at 6/2021/3, and other field management measures are managed according to local high-yield cultivation. The winter wheat has no diseases, weeds, good field environment, no drought, waterlogging and the like in the whole growth period.
As shown in fig. 1: the invention relates to a method for monitoring space-time variation of winter wheat SPAD, which comprises the following steps:
step S001: unmanned aerial vehicle-based data collection;
step S011: 11 on the day of fertility (wheat heading period 2021, month 18, flowering period 2021, month 29, and post-grouting 2021, month 5, 24) using a Phantom 4Multispectral RTK (DJI Technology co., shenzhen, china) (P4M) drone: 00 to 13:00 collecting Multispectral (MS) data;
as an embodiment provided by the present invention, P4M preferably has 5 single color sensors of 208 ten thousand pixels: blue, green, red, red_edge and NIR, the central wavelength of which is 450nm,560nm,650nm,730nm and 840nm, the bandwidths are + -16 nm, + -26 nm, and real-time kinetic (RTK) system is equipped at the same time, the RTK can realize the vertical positioning precision + -1.5 cm and the horizontal positioning precision + -1 cm, so that P4M can obtain the spectrum and texture information with higher precision.
As an embodiment provided by the present invention, preferably, to avoid missing image spectrum and texture information of the unmanned aerial vehicle due to cloud coverage, the method comprises the following steps of 11:00 to 13:00, selecting a weather condition with clear sky, no wind and no cloud and stable sunshine radiation intensity for flying.
As an embodiment provided by the invention, the ground preferably controls 4 radiometric calibration plates of known reflectivity for subsequent data processing to develop radiometric calibration.
As an embodiment provided by the present invention, it is preferable that the UAV equipped with the MS sensor automatically adjusts exposure according to the environment in flight in order to improve the accuracy of the image. The DJI GS PRO software (https:// www.dji.com/cn/group-station-PRO /) is adopted to plan the flight route in advance, a predefined flight plan is developed by depending on an unmanned aerial vehicle automatic driving system, each flight lasts for 20 minutes, the flight height is 30m, the flight speed is 2m/s, the heading and side overlap degree are respectively 90% and 85%, and the image resolution is 1600 multiplied by 1300 pixels.
Step S002: collecting field data;
step S021: the field measurements for each growth period were performed prior to the flight of the unmanned aerial vehicle on the same day, and nondestructive winter wheat SPAD information was determined in 36 plots using SPAD-502 chlorophyll machine (Konica Minolta Optics inc., osaka, japan). 3 winter wheat plants with average growth vigor are selected from each plot, SPAD readings at positions of 1/6, 3/6 and 5/6 of the length of wheat flag leaves are measured, the average value of 9 SPAD readings at a sampling place is used as an actual measurement value of SPAD of the plot, 324 times of the average value is collected in each growing period, and the actual measurement data of the SPAD of 36 plots are calculated;
step S003: unmanned aerial vehicle image preprocessing, importing unmanned aerial vehicle MS multispectral images acquired in each growth period into PIX4 Dmoper software (4.4.12version,Pix4D SA,Prilly,Switzerland) for image stitching, aligning the images by utilizing a characteristic point matching algorithm, and acquiring dense point clouds and texture grids based on unmanned aerial vehicle images and position data. In order to improve the data quality, the invention uses the ground radiation calibration plate image information with known reflectivity to carry out radiation correction on the MS image based on an empirical linear method. Creating a shape file by using ArcGIS10.2 (Environmental Systems Research Institute, inc, redLands, CA, USA), dividing a sample area according to an orthophoto map of a test area, generating 36 samples superimposed on a wheat image, and giving ID information one by one;
specifically, field measurements for each growth period were performed prior to unmanned aerial vehicle flight on the same day, and nondestructive winter wheat SPAD information was determined in 36 plots using SPAD-502 chlorophyll machines (Konica Minolta Optics inc., osaka, japan). 3 winter wheat plants with average growth vigor are selected from each plot, the SPAD readings at positions 1/6, 3/6 and 5/6 of the length of wheat flag leaves are measured, the average value of 9 SPAD readings at the sampling place is used as the measured value of SPAD of the plot, 324 times of SPAD measurements are collected in each growth period, and the measured SPAD data of 36 plots are calculated.
Step S031: spectral feature extraction: the invention takes the vegetation index as the spectrum characteristic (SF), and obtains stronger vegetation information factors by the characteristic wave bands through operation or combination, thereby increasing the expression capability of remote sensing data to a certain extent. Acquiring a plurality of 5 original wave band reflectivities of a winter wheat key growth period (heading period, flowering period and grouting later period) canopy from an MS multispectral image;
step S032: texture feature extraction: texture is a common means for representing image information, and reflects important information such as surface structure and spatial arrangement in an image without depending on brightness, and generally detects the periodicity of the image by using high-energy narrow peaks in a spectrum. According to the invention, texture information of blue, green, red edges and near infrared bands in MS multispectral images of winter wheat canopy is extracted by adopting a Grayscale co-occurrence matrix (GLCM) method;
step S004: the average number of SPAD values per plot was calculated and analyzed for direct characterization of winter wheat growth. Then merging spectrum textures of the spectrum features, the texture features and the image information and respectively constructing SPAD prediction models;
step S005: winter wheat SPAD information was predicted using support vector machine regression (SVR) algorithm based on R language version.1.3 (R Foundation, vienna, austria).
Unmanned aerial vehicle includes the monochromatic sensor of 5 208 ten thousand pixels, and monochromatic sensor is respectively: blue, green, red, red edge and NIR, the central wavelength of monochromatic sensor is 450nm,560nm,650nm,730nm and 840nm respectively, and monochromatic sensor's bandwidth is 16nm, 26nm respectively, unmanned aerial vehicle includes real-time kinetic (RTK) system, real-time kinetic (RTK) system's vertical precision and horizontal precision are ± 0.1M respectively for P4M can acquire spectrum and texture information that the precision is higher.
In the step S003, the image stitching method includes the following steps:
step S0311: aligning the histogram by utilizing a characteristic point matching algorithm, and acquiring a dense point cloud and a texture grid based on the unmanned aerial vehicle image and the position data;
step S0312: correcting the difference between the wave bands to improve the data quality;
step S0313: a shapefile is created by ArcGIS10.2 (Environmental Systems Research Institute, inc, redLands, CA, USA), a plot area is divided according to an orthophoto map of a test area, 36 units superimposed on a winter wheat image are generated, and ID information is given to the units one by one.
In the step S0312, the method for correcting the difference between the bands is as follows:
distributing and controlling 4 radiation calibration plates with known reflectivity on the ground;
the MS multispectral image is subjected to radiation correction by using ground radiation calibration plate image information with known reflectivity based on an empirical linear method.
In the step S031, preferably, the reflectance of the wave band and the vegetation index are used as Spectral Features (SF), and the vegetation index is obtained by calculating or combining characteristic wave bands, so that the expression capability of remote sensing data is increased to a certain extent; acquiring 5 original band reflectivities of winter wheat canopy in heading period, flowering period and post-grouting period from an MS image for calculating a vegetation index; in order to study the influence of nitrogen level and growth period on leaf chlorophyll-based SPAD values, 30 spectral characteristics widely used for crop growth monitoring and parameter evaluation are selected, and the spectral characteristics are divided into 5 categories, namely 5 original wave bands (I), 7 vegetation indexes (II) which are only composed of visible light wave bands, 5 vegetation indexes (III) which are composed of red side wave bands and do not exist in a near infrared wave band, 8 vegetation Indexes (IV) which are composed of near infrared wave bands and do not exist in a red side wave band, and 5 vegetation indexes (V) which are composed of near infrared wave bands and red side wave bands together, and are shown in a table 1.
As an embodiment of the present invention, preferably, in the step S032, the texture information includes eight indexes, such as Mean (Me), variance (Va), homogeneity (Ho), contrast (Cn), dissimilarity (Di), entropy (En), second movement (Se), and Correlation (Cr), as shown in table 2. Texture reflects important information such as surface structure and spatial arrangement in an image independent of brightness, and generally detects periodicity of the image by using high-energy narrow peaks in a spectrum. The invention adopts a Grey Level Co-occurrence Matrices (GLCM) method to extract texture information of blue, green, red and near infrared bands in winter wheat canopy MS images.
As an embodiment provided by the invention, preferably, the feature selection technology cannot be limited to only considering the relation between the features and the SPAD, but also should consider the relation between the features, and the multispectral remote sensing data of the unmanned aerial vehicle is processed to determine the optimal method for maximally improving the SPAD estimation in the late growth stage. And (3) respectively optimizing spectral features and texture features which are sensitive to the SPAD by adopting a feature selection strategy, further constructing a SPAD prediction model to check the difference between the feature selection strategy and the feature selection strategy, and fusing different types of feature subsets together by using a feature fusion strategy to construct a SPAD estimation model.
Table 1 spectral characteristics
TABLE 2 texture characterization
In step S004, when the SPAD prediction model is constructed, a RFE (Recursive feature elimination) algorithm and a Boruta algorithm are adopted to screen different feature parameter sets, and the feature parameter sets are compared with the model performance of the feature sets which are not screened to obtain the optimal feature combination. Specifically, the Recursive Feature Elimination (RFE) algorithm first uses all feature training models, calculates the importance of each feature and ranks it, uses each feature subset training model, compares the model results obtained for each subset, retrains the model based on the particular feature, and repeats this process until the optimal feature combination is screened to obtain the maximized model performance. RFE may operate based on a variety of algorithmic models, such as the usual machine learning methods of RF and SVR. In the invention, the RFE is operated based on the RF algorithm, and the optimal feature set is output after the operation is finished;
the Boruta algorithm is a wrapper based on a random forest algorithm, is an integrated method for classifying a plurality of independent decision tree votes, classifies all trees based on given attributes, and calculates the importance of all trees, namely a Z score reflecting the precision fluctuation among the trees in the forest. In operation, boruta creates a "shadow" attribute obtained by reshuffling the original attribute and randomly breaks the order of feature parameters, and when calculating feature importance, the feature parameters are classified into three categories, namely, features with Z score significantly higher than the "shadow" attribute are called fixed (important), features with Z score close to the "shadow" attribute are called fixed (possibly important), and features with Z score significantly lower than the "shadow" attribute are called reject (unimportant). The method divides the Borata algorithm screening result into two types of Confirmed and Tentative, analyzes the influence of two characteristic parameter sets on the construction model and compares the precision of the two characteristic parameter sets, has the advantage of random forest algorithm, has low running time cost, and can run the result without depending on parameter adjustment.
The characteristic parameter set variable screened by the RFE algorithm takes R-as prefix, the characteristic parameter set screened by the Borata algorithm (the characteristic parameter set is of two types of Confirmed and Confirmed+Tentative) is respectively expressed by C-and CT-, and for example, the spectrum characteristic screened by the RFE algorithm is expressed by R-SF;
the characteristic parameter set screened by the Borata algorithm comprises two types of Confirmed and Confirmed+Tentative.
As an embodiment of the present invention, preferably, in the step S004, when the SPAD prediction model is constructed, feature fusion is a method of constructing the model by fusing different types of remote sensing features together. Fusing the features into a feature selection policyAfter slightly combined execution, the stability, the precision and the robustness of the winter wheat SPAD estimation model are all in the leading position, and the precision improvement amplitude of the model gradually becomes larger along with the promotion of the growth period. And the model precision in the later stage of grouting is improved to the highest extent, and R is compared with a model using only initial spectral features or initial texture features 2 Val The RMSE is improved by 0.092 to 0.202 Val Reduced by 0.076 to 4.916, RPD Val Lifting by 0.237 to 0.960; in the present invention, the feature fusion strategy is mainly divided into 2 parts:
1. firstly, fusing spectral features and texture features, constructing a SPAD estimation model after the feature selection method is optimized, and comparing the SPAD estimation model with a model constructed by fused features which do not participate in feature selection;
2. based on the feature selection strategy, the SPAD estimation model is built by fusing the optimized feature subsets of different categories so as to seek the SPAD estimation model with optimal performance in the late growth period of wheat.
Embodiment two:
as an embodiment provided by the present invention, more preferably, the method for predicting chlorophyll content of winter wheat based on SPAD value comprises the following steps:
step S0051: dividing the SPAD characteristic parameter set of winter wheat in each growth period into a training set and a verification set, randomly sampling the characteristic parameter set, taking 2/3 of the characteristic parameter set for model training, and 1/3 of the characteristic parameter set for model verification;
step S0052: 3 statistical indexes are adopted to test a machine learning model, and the 3 statistical indexes are respectively: the coefficient of determination (R) 2 ) Determining coefficients, root Mean Square Error (RMSE) root mean square error, the ratio of performance to deviation (RPD) performance to bias ratio; wherein:
in the method, in the process of the invention,and y i For the observations and measured values of SPAD, +.>The average value of the SPAD observation values is represented by n, the number of samples is represented by the number of samples, and the standard deviation of the observation values is represented by SD; descriptive statistics for SPADs of calibration and validation sets are shown in table 3;
TABLE 3 Table 3
As an embodiment provided in the present invention, more preferably, in order to obtain more texture information for feature selection, a maximum value (MAX), a minimum value (MIN), a MEAN value (MEAN), and a Standard Deviation (SD) of each sample region GLCM index are calculated, resulting in 5 (bases) ×8 (GLCM indices) ×4 (statistical metrics) =160 texture indexes in total.
A winter wheat SPAD space-time change monitoring method comprises the steps of acquiring a winter wheat canopy high-resolution image through a multispectral sensor carried by an unmanned aerial vehicle, extracting spectral features and texture features of wheat by using an image processing technology, excavating the spectral features and the texture features of winter wheat in the later growth stage by using two feature selection methods, and developing a winter wheat SPAD monitoring model based on the combination of a feature selection strategy and a feature fusion strategy so as to solve the following problems:
(1) Defining spectral characteristics and texture characteristics of the winter wheat sensitive to SPAD in the reproductive growth stage;
(2) Evaluating the performance of a winter wheat SPAD prediction model under a feature selection strategy;
(3) And (3) exploring a characteristic selection strategy and a characteristic fusion strategy to estimate the potential of SPAD in the late growth stage of winter wheat.
Based on the first and second embodiments, the spectral characteristics of the winter wheat canopy are closely related to the SPAD, fig. 2 shows the spectral characteristics of the winter wheat in three growth periods and the space-time variation of the SPAD, and the degree of the sample plot color intuitively shows the variation of the spectral characteristics NDVI, GCVI and CVI; NDVI, GCVI, CVI the spectral characteristics a, b and c of the plot represent the heading, flowering and post-grouting phases of winter wheat, respectively. Fig. 3 is a graph of wheat spectral characteristics versus SPAD fold line variation at 4 different nitrogen application levels, with nodes representing averages of the characteristics. The spectral characteristics and SPAD all show obvious differences under different nitrogen gradient treatments, and the overall trend of the SPAD gradually rises with the increase of the nitrogen application amount. With the promotion of the growth period, the spectral characteristic changes of Huaimai 44, ningnong 999 and Ningmai 13 are consistent with the overall trend of SPAD dynamic change, and no obvious difference exists;
based on the first and second examples, the spatial and temporal changes of the texture features and SPAD of the three different growth periods are shown in fig. 4 (fig. 4, the spatial and temporal changes of the canopy texture features and SPAD of winter wheat: MEAN.R.Di, MEAN.NIR.Me, MEAN.NIR.Cr represents the texture features a, b and c of the plot, respectively, represent the heading, flowering and post-grouting periods of winter wheat: MEAN.NIR.Di, MEAN.NIR.Me is consistent with the trend of the spectral features, and the trend of decrease is shown from heading to post-grouting period, whereas mean.r.cr is the opposite, as shown in fig. 5 (4 plots of the texture features of wheat with nitrogen application levels versus SPAD fold line, nodes represent the average of the features), MEAN.NIR.Di, MEAN.NIR.Me generally shows a trend of decrease after increase with increasing nitrogen application amount, and peaks are generally of N2 level and N1 level.
Based on the first and second embodiments, it can be obtained that 30 spectral features and 160 texture features are extracted from the multispectral image of the unmanned aerial vehicle in the present invention, not all features contribute to wheat SPAD estimation, and redundant features may affect model accuracy. According to a wheat spectral feature screening radar chart (obtained by experiments and not provided at first) in three periods, 3 feature sets selected by the spectral features in the heading period are the same, and 5 spectral feature types are selected; the spectral characteristics of the types IV and V are selected by C-SF and CT-SF, and the R-SF mainly selects the type IV as an optimal variable set; the best performance of type iv and type v in the later stage of grouting, the 3 feature set preferred spectral feature type trends appear highly consistent. The spectral feature types iv and v are common in that they are each composed of near infrared band participation (spectral features are divided into 5 types, respectively, 5 original bands (i), 7 vegetation indexes (ii) composed of only visible light bands, 5 vegetation indexes (iii) composed of red side bands and no near infrared band, 8 vegetation indexes (iv) composed of near infrared bands and no red side band, and 5 vegetation indexes (v) composed of near infrared and red side bands together), as shown in fig. 6, the optimal variable numbers of the Boruta and RFE feature selection methods are selected for different feature sets in different breeding periods. A. B and C represent the heading stage, the flowering stage and the post-grouting stage of winter wheat. SF represents spectral features, R-SF is a spectral feature dataset retained by RFE method screening, C-SF is a "fixed" spectral feature dataset retained by Borata method screening, and CT-SF is a "fixed+Tentative" spectral feature dataset retained by Borata method screening. TF represents texture features, R-TF, C-TF and CT-TF are as above. According to the three-period texture characteristic wave band type screening result in the invention, the ear-picking period Red, the Rededge and the NIR wave band texture performance are better and far greater than Blue and Green wave bands. The grain characteristic wave band types of the three screening results in the flowering period have consistent trend, and the Red wave band and the N IR wave band have better performance; the texture characteristic wave bands of the three screening results in the later period of grouting are consistent, and the types of the 5 texture characteristic wave bands have no obvious difference. The Red and N IR wave bands are stable in texture performance, the feature retention number after screening except the later period of grouting is obviously higher than that of other wave bands, and the 5 wave bands at the later period of grouting are consistent in performance and have no obvious difference.
Comparative example:
1. the method is characterized in that SPAD estimation models of heading period, flowering period and grouting later period are respectively established based on spectral features and texture features extracted from the multispectral images of the unmanned aerial vehicle in combination with feature selection strategies, an original feature set is added in each stage to perform model construction, and 24 (4 x 2 x 3) SPAD regression prediction models are constructed by utilizing 4 feature sets, 2 feature types (spectral features and texture features) and 3 growth periods.
The filtered dataset model accuracy is almost higher than the initial dataset accuracy in terms of model accuracy before and after feature selection, with some exceptions. For example flowering period SF (verification set: R) 2 =0.805, rmse=3.211, rpd=1.998) is higher than the accuracy of the R-SF, C-SF, CT-SF data set to construct the SPAD estimation model. The 3 data sets formed after feature selection have different precision, and are inconsistent in different feature types of different growth periods, such as R-TF data set precision (verification set: R 2 =0.800, rmse=3.732, rpd=2.080) below C-TF, CT-TF, whereas the flowering phase R-TF dataset accuracy (validation set: r is R 2 =0.799, rmse=2.941, rpd= 2.181) is higher than C-TF, CT-TF. The model accuracy of the C-dataset and the CT-dataset of the filter division of the Borata algorithm is not constant.
2. And executing a feature selection strategy after applying a feature fusion strategy based on the spectral features and the texture features extracted from the multispectral images of the unmanned aerial vehicle, and respectively establishing SPAD estimation models of the heading period, the flowering period and the post grouting period. The original feature set is added in each stage for model construction, and 12 (4*3) SPAD regression prediction models are constructed by utilizing 4 feature sets, 1 feature type (fusion of spectral features and texture features) and 3 growth periods.
Under the characteristic strategy, the accuracy of a winter wheat heading stage canopy SPAD estimation model (R-SFTF) constructed by combining an RFE characteristic selection method with an SFTF data set is optimal. The performance index of the model is as follows: verification set: r is R 2 =0.861, rmse=3.604, rpd=2.154. The winter wheat flowering period canopy SPAD estimation model (C-SFTF) constructed by combining the Borata feature selection method with the SFTF data set has optimal precision, and specific performance indexes are as follows: verification set: r is R 2 =0.740, rmse=3.256, rpd=1.971. Winter wheat grouting later-stage canopy SPAD estimation constructed by combining RFE feature selection method with SFTF data setModel (R-SFTF) performs best, where the validation set: r is R 2 ＝0.761，RMSE＝6.250，RPD＝1.983。
3. And executing a feature fusion strategy after applying a feature selection strategy based on the spectral features and the texture features extracted from the multispectral images of the unmanned aerial vehicle, and respectively establishing SPAD estimation models of the heading period, the flowering period and the post grouting period. Each stage utilizes 9 (3*3) feature sets, 1 feature type (feature fusion set) and 3 growth periods to construct 27 (9*3) SPAD predictive models, and an unfiltered SFTF feature set is added at each stage as a comparison for a total of 30 regression models.
Under the strategy, since the heading period R-SF, C-SF and CT-SF feature sets are the same, 3 optimal estimation models which are fused with R-TF data are respectively provided in the period, wherein the optimal estimation models are R-SF-R-TF, C-SF-R-TF and CT-SF-R-TF, and the performance indexes of the models are as follows: verification set: r is R 2 =0.857, rmse=3.134, rpd= 2.477. The accuracy of a winter wheat flowering period canopy SPAD estimation model (R-SF-R-TF) constructed by combining an RFE feature selection method with SF and TF data sets is optimal, and specific performance indexes are as follows: verification set: r is R 2 =0.807, rmse=2.850, rpd= 2.251. The performance of a winter wheat grouting later-stage canopy SPAD estimation model (C-SF-CT-TF) constructed by combining a Borata feature selection method with an SF and TF data set is optimal, wherein a validation set is formed: r is R 2 ＝0.809，RMSE＝5.878，RPD＝2.108。
Thus, to further analyze SPAD estimation model accuracy, fig. 7 shows a scatter plot of measured and predicted values for the optimal model at heading, flowering and post-grouting phases of the 63 models built in total in accordance with the present invention. The graph shows that the optimal monitoring model in the late growth stage of winter wheat is a SPAD estimation model constructed under a dual-feature strategy, data points in each period are gathered near a 1:1 line, the predicted value output by the model has good agreement with the actual measured value in the field, and the error is small. The results show that the models can accurately estimate the winter wheat canopy SPAD.
Four different statistical indexes of each sample area based on GLCM features are extracted through further analysis of the statistical indexes of the texture features: MEAN (MEAN), standard Deviation (SD), maximum (MAX), and Minimum (MIN). The common MEAN index MEAN feature retention is always kept leading in the four statistical indexes, but in the importance sorting, the average MEAN index MEAN feature retention shows a trend of gradually decreasing importance along with the progress of the growth period. The change trend of the importance of other indexes is opposite to MEAN, the importance gradually rises from the heading period to the later period of grouting, and the importance reaches a peak value in the later period of grouting, wherein the SD is the best in performance trend, which shows that other statistical indexes also have the potential of estimating the wheat SPAD, in particular SD. The result of the invention shows that the SD information index of the texture features also has the potential of estimating the SPAD of crops.
In the description of the present specification, the descriptions of the terms "one embodiment," "example," "specific example," and the like, mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The preferred embodiments of the invention disclosed above are intended only to assist in the explanation of the invention. The preferred embodiments are not exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. The invention is limited only by the claims and the full scope and equivalents thereof.
Claims (9)
1. The method for monitoring the space-time variation of the winter wheat SPAD is characterized by comprising the following steps of:
step S001: collecting multispectral data of winter wheat in a growing period by using an unmanned aerial vehicle;
step S002: determining the relative chlorophyll content value of winter wheat in a plurality of sample areas by utilizing a SPAD-502 type chlorophyll meter;
step S003: unmanned aerial vehicle image preprocessing: the multispectral images of the unmanned aerial vehicle collected in each growth period are imported into PIX4 Dapp er software for image stitching;
step S004: calculating the average number of the relative chlorophyll content values of each land block, analyzing the average number, and respectively constructing a SPAD prediction model by using the spectrum characteristics, the texture characteristics and the spectrum texture characteristic fusion set;
step S005: and predicting the chlorophyll content of winter wheat based on the SPAD value by adopting a support vector machine regression algorithm based on the R language.
2. The method for monitoring the space-time variation of winter wheat SPAD according to claim 1, wherein the unmanned aerial vehicle comprises 5 single-color sensors with 1208 ten thousand pixels, and the single-color sensors are respectively: the central wavelengths of the monochromatic sensors are 450nm,560nm,650nm,730nm and 840nm respectively, the bandwidths of the monochromatic sensors are +/-16 nm, +/-16 nm and +/-26 nm respectively, the unmanned aerial vehicle comprises an RTK system, and the vertical precision of the RTK system is +/-1.5 cm and the horizontal precision of the RTK system is +/-1 cm.
3. The method for monitoring space-time variation of winter wheat SPAD according to claim 1, wherein in the step S003, the method for image stitching comprises the steps of:
step S0311: aligning the histogram by utilizing a characteristic point matching algorithm, and acquiring a dense point cloud and a texture grid based on the unmanned aerial vehicle image and the position data;
step S0312: correcting the difference between the wave bands;
step S0313: and creating a vector file by using ArcGIS10.2, dividing a sample area according to an orthophoto map of the test area, generating a plurality of vector cells overlapped on the winter wheat image, and giving ID information to the vector cells one by one.
4. The method for monitoring the space-time variation of the SPAD of winter wheat according to claim 1, wherein said step S003 further comprises the steps of:
step S031: spectral feature extraction: acquiring a plurality of original wave band reflectivities of a winter wheat growing period canopy from a multispectral image;
step S032: texture feature extraction: and extracting texture information of blue, green, red edges and near infrared bands in the winter wheat canopy multispectral image by adopting a gray level co-occurrence matrix method.
5. A method for monitoring space-time variation of winter wheat SPAD according to claim 3, wherein in the step S0312, the method for correcting the difference between the wave bands is as follows:
a plurality of radiation calibration plates with known reflectivity are distributed and controlled on the ground;
the multispectral image is subjected to radiation correction by using ground radiation calibration plate image information with known reflectivity based on an empirical linear method.
6. The method according to claim 4, wherein in the step S031, 30 spectral features for crop growth monitoring and parameter evaluation are selected, and the spectral features are classified into 5 categories, namely, 5 original bands, 7 vegetation indexes consisting of only visible bands, 5 vegetation indexes consisting of red bands and no near infrared bands, 8 vegetation indexes consisting of near infrared bands and no red bands, and 5 vegetation indexes consisting of both near infrared and red bands.
7. The method of claim 4, wherein in step S032, the texture information includes texture information including a mean, variance, homogeneity, contrast, variability, entropy, angular second moment and correlation, and the maximum value, minimum value, mean and standard deviation of each texture index are calculated to obtain 5×8×4 texture indexes.
8. The method for monitoring the space-time variation of the SPAD of the winter wheat according to claim 1, wherein in the step S004, when the SPAD prediction model is constructed, different characteristic parameter sets are screened by adopting an RFE algorithm and a Boruta algorithm, and the optimal characteristic combination is obtained by fusing the different characteristic parameter sets and comparing and analyzing the model performances of the characteristic parameter sets which are not screened;
the feature parameter set variable screened by the RFE algorithm takes R-as prefix, and the feature parameter set screened by the Borata algorithm is respectively expressed by C-and CT-;
the feature parameter set screened by the Borata algorithm comprises two types of 'confirmed' and 'pending'.
9. The method for monitoring the SPAD space-time variation of winter wheat according to claim 7, wherein in the step S005, the method for predicting the chlorophyll content of winter wheat based on SPAD values comprises the following steps:
step S0051: dividing the SPAD characteristic parameter set of winter wheat in each growth period into a training set and a verification set;
step S0052: 3 statistical indexes are adopted to test a machine learning model, and the 3 statistical indexes are respectively: determining the coefficient R 2 Root mean square error RMSE and performance to bias ratio RPD; wherein:
in the method, in the process of the invention,and y i For the observations and measured values of SPAD, +.>The average value of the SPAD observations is represented by n, the number of samples, and SD, the standard deviation of the observations.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311794390.XA CN117760984A (en) | 2023-12-25 | 2023-12-25 | Winter wheat SPAD space-time change monitoring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311794390.XA CN117760984A (en) | 2023-12-25 | 2023-12-25 | Winter wheat SPAD space-time change monitoring method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117760984A true CN117760984A (en) | 2024-03-26 |
Family
ID=90317894
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311794390.XA Pending CN117760984A (en) | 2023-12-25 | 2023-12-25 | Winter wheat SPAD space-time change monitoring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117760984A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107330892A (en) * | 2017-07-24 | 2017-11-07 | 内蒙古工业大学 | A kind of sunflower disease recognition method based on random forest method |
| CN108760660A (en) * | 2018-08-17 | 2018-11-06 | 山东农业大学 | A kind of period of seedling establishment leaves of winter wheat chlorophyll contents evaluation method |
| CN110082309A (en) * | 2019-05-20 | 2019-08-02 | 中国水利水电科学研究院 | Canopy of winter wheat SPAD value integrated spectral monitoring model method for building up |
| CN110567891A (en) * | 2019-09-16 | 2019-12-13 | 中国水利水电科学研究院 | Winter wheat canopy chlorophyll estimation system and method |
| CN115512123A (en) * | 2022-10-19 | 2022-12-23 | 上海第二工业大学 | Multi-period key growth characteristic extraction and time period classification method for hypsizygus marmoreus |
| CN116485757A (en) * | 2023-04-25 | 2023-07-25 | 江苏华源节水股份有限公司 | Winter wheat total nitrogen content prediction method |
-
2023
- 2023-12-25 CN CN202311794390.XA patent/CN117760984A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107330892A (en) * | 2017-07-24 | 2017-11-07 | 内蒙古工业大学 | A kind of sunflower disease recognition method based on random forest method |
| CN108760660A (en) * | 2018-08-17 | 2018-11-06 | 山东农业大学 | A kind of period of seedling establishment leaves of winter wheat chlorophyll contents evaluation method |
| CN110082309A (en) * | 2019-05-20 | 2019-08-02 | 中国水利水电科学研究院 | Canopy of winter wheat SPAD value integrated spectral monitoring model method for building up |
| CN110567891A (en) * | 2019-09-16 | 2019-12-13 | 中国水利水电科学研究院 | Winter wheat canopy chlorophyll estimation system and method |
| CN115512123A (en) * | 2022-10-19 | 2022-12-23 | 上海第二工业大学 | Multi-period key growth characteristic extraction and time period classification method for hypsizygus marmoreus |
| CN116485757A (en) * | 2023-04-25 | 2023-07-25 | 江苏华源节水股份有限公司 | Winter wheat total nitrogen content prediction method |
Non-Patent Citations (4)
| Title |
|---|
| QUAN YIN等: "Better Inversion of Wheat Canopy SPAD Values before Heading Stage Using Spectral and Texture Indices Based on UAV Multispectral Imagery", REMOTE SENSING, vol. 15, no. 20, 12 October 2023 (2023-10-12), pages 4 - 20 * |
| YUWEI WANG等: "Estimating Relative Chlorophyll Content in Rice Leaves UsingUnmanned Aerial Vehicle Multi-Spectral Images and Spectral–Textural Analysis", AGRONOMY, vol. 13, no. 6, 1 June 2023 (2023-06-01), pages 1 - 17 * |
| 王建荣: "基于多特征融合的无人机航拍图像识别研究", 中国优秀硕士学位论文全文数据库信息科技辑, no. 04, 15 April 2016 (2016-04-15), pages 40 * |
| 苏祥祥等: "基于无人机多光谱遥感的冬小麦生长后期 SPAD 变化监测", 第二十届中国作物学会学术年会论文摘要集, 30 November 2023 (2023-11-30) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Orchard management with small unmanned aerial vehicles: A survey of sensing and analysis approaches | |
| Ballesteros et al. | Vineyard yield estimation by combining remote sensing, computer vision and artificial neural network techniques | |
| CN106372592B (en) | A kind of winter wheat planting area calculation method based on winter wheat area index | |
| US20230292647A1 (en) | System and Method for Crop Monitoring | |
| CN107273820B (en) | Remote sensing classification method and system for land coverage | |
| CN111241912A (en) | Multi-vegetation index rice yield estimation method based on machine learning algorithm | |
| CN108458978B (en) | Sensitive waveband and waveband combination optimal tree species multispectral remote sensing identification method | |
| CN114821362B (en) | Multi-source data-based rice planting area extraction method | |
| CN112557393B (en) | Wheat leaf layer nitrogen content estimation method based on hyperspectral image fusion map features | |
| Liang et al. | Improved estimation of aboveground biomass in rubber plantations by fusing spectral and textural information from UAV-based RGB imagery | |
| CN112287886B (en) | Wheat plant nitrogen content estimation method based on hyperspectral image fusion map features | |
| CN109784300A (en) | A kind of crops science survey production method and system | |
| CN116310840A (en) | Winter wheat remote sensing identification method integrating multiple key weather period spectral features | |
| Liu et al. | UAV multispectral images for accurate estimation of the maize LAI considering the effect of soil background | |
| Yuan et al. | Research on rice leaf area index estimation based on fusion of texture and spectral information | |
| CN117760984A (en) | Winter wheat SPAD space-time change monitoring method | |
| Bao et al. | Forecasting spring maize yield using vegetation indices and crop phenology metrics from UAV observations | |
| CN114154694A (en) | Method for predicting plant growth state in cotton topping period based on multispectral remote sensing of unmanned aerial vehicle | |
| CN113514402A (en) | System and method for predicting chlorophyll content of winter wheat | |
| Rananavare et al. | Monocot Crop Yield Prediction using Sentinal-2 Satellite Data | |
| Luo et al. | Staple crop mapping with Chinese GaoFen-1 and GaoFen-6 satellite images: A case study in Yanshou County, Heilongjiang Province, China | |
| KR20220094512A (en) | A method for monitoring crop growth by analyzing spectroscopic images through remote sensing | |
| ISHOLA | APPLICATION OF SATELLITE BASED REMOTE SENSING TO THE ESTIMATION AND MONITORING OF CROP HEALTH | |
| Martello | Orbital, aerial, and proximal sensing applied to monitoring the spatial variability of coffee plantations | |
| Hall | Defining grapevine and vineyard characteristics from high spatial resolution remotely sensed optical imagery |
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 |