CN110658211A - Method for extracting spectral characteristics of cotton canopy of aphid pests in cotton bud period and estimating grade - Google Patents
Method for extracting spectral characteristics of cotton canopy of aphid pests in cotton bud period and estimating grade Download PDFInfo
- Publication number
- CN110658211A CN110658211A CN201911087521.4A CN201911087521A CN110658211A CN 110658211 A CN110658211 A CN 110658211A CN 201911087521 A CN201911087521 A CN 201911087521A CN 110658211 A CN110658211 A CN 110658211A
- Authority
- CN
- China
- Prior art keywords
- cotton
- aphid
- canopy
- spectrum
- spectral
- 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
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a method for extracting spectral characteristics and estimating the grade of a cotton canopy of aphid pests in a cotton bud period, which comprises the following steps: firstly, acquiring a hyperspectral image of a cotton canopy of aphid damage in a cotton bud period by using a low-altitude unmanned aerial vehicle carried imaging high-spectrum spectrometer; secondly, analyzing canopy spectral response differences of cotton stressed by different aphid levels; thirdly, screening out sensitive spectrum bands of the cotton canopy of the cotton aphid damage in the bud period by using a ratio derivative method according to the hyperspectral image of the cotton canopy of the cotton aphid damage in the bud period; fourthly, constructing an aphid level estimation model based on a partial least square method of the sensitive spectrum ratio derivative spectrum value; fifthly, applying the model to an imaging hyperspectral image to obtain an aphid grade distribution map of a field scale; the method for extracting the spectral characteristics of the cotton canopy of the aphid pests at the bud period of the cotton and estimating the level is used for rapidly monitoring the level distribution of the aphid pests at the field scale, is convenient for pesticide application and monitoring of the aphid pests of the cotton, can assist in accurate and quantitative pesticide application, reduces environmental pollution, and has very important significance for actual production.
Description
Technical Field
The invention belongs to the technical field of cotton planting and maintenance, and particularly relates to a method for extracting spectral characteristics and estimating a grade of a cotton canopy of aphid pests in a cotton bud period.
Background
The cotton aphids are one of the main insect pests affecting the cotton yield, have large harm degree and wide occurrence range, are the global cotton pests and are distributed in various large cotton areas at home and abroad. The cotton aphids are withered by directly sucking the juice of the cotton plants, and honeydew secreted by the cotton aphids influences the normal photosynthesis and physiological action of cotton, pollutes cotton fibers and induces the parasitism of mold fungi. The occurrence of cotton aphids in large areas can cause serious yield loss in cotton fields.
Although the traditional field investigation mode of crop diseases and insect pests can obtain better investigation effect, the traditional field investigation mode wastes time and labor, has poor timeliness and is not suitable for large areas. The monitoring of crop diseases and insect pests by using the characteristics of rapidness, dynamics and no damage of a remote sensing technology becomes a research hotspot in the field of current agricultural remote sensing. Some scholars monitor the growth condition of the cotton after the growth of the cotton is stressed by utilizing near-earth hyperspectral data at the leaf scale and the canopy scale, and research results show that the spectral characteristics of the cotton subjected to different stress degrees are different, so that a foundation is laid for monitoring the grade of the cotton insect damage. However, the near-earth hyperspectral data is accurate, but has spatial randomness, and cannot reflect the spatial characteristics of pest and disease occurrence.
In recent years, many scholars begin to apply the unmanned aerial vehicle to farmland ecological environment information monitoring and acquisition due to the advantages of flexible data measurement time, high space-time resolution, large observation range and the like. The sensors carried by the agricultural unmanned aerial vehicle are divided into two types, one type is that a common digital camera and a multispectral camera are used as main sensors, the space resolution of the obtained image data is high, the processing is quick, but the wave band is few, the spectral information is limited, and the defects limit the application of the remote sensing technology in the monitoring of crop diseases and insect pests; the other type is an imaging hyperspectral meter, the acquired data has the advantages of high spatial resolution and high spectral resolution, the defects of a common digital camera and a multispectral camera can be well overcome, the advantages of very strong capability in extracting a spatial distribution range of plant diseases and insect pests are achieved, the data volume is overlarge, the processing time is long, the requirements on computer hardware used for processing the data are high, and the imaging hyperspectral meter is difficult to be applied to actual production. Therefore, a small number of wave bands which are most sensitive to plant stress information are selected, a pest grade estimation model is constructed, and the method has very important significance for improving the processing efficiency of the imaging hyperspectral data of the unmanned aerial vehicle and applying the method to actual production.
Disclosure of Invention
The invention mainly aims to solve the defects of the prior art, extract the sensitive spectrum band of the canopy of the aphid cotton in the bud period and construct an aphid grade estimation model based on the sensitive spectrum.
In order to achieve the purpose, the invention provides the following technical scheme: the method for extracting the spectral characteristics of the cotton canopy of the aphid pests in the bud period of the cotton and estimating the grade comprises the following steps: firstly, acquiring a hyperspectral image of a cotton canopy of aphid damage in a cotton bud period by using a low-altitude unmanned aerial vehicle carried imaging high-spectrum spectrometer; secondly, analyzing canopy spectral response differences of cotton stressed by different aphid levels; thirdly, screening out sensitive spectrum bands of the cotton canopy of the cotton aphid damage in the bud period by using a ratio derivative method according to the hyperspectral image of the cotton canopy of the cotton aphid damage in the bud period; fourthly, constructing an aphid level estimation model based on a partial least square method of the sensitive spectrum ratio derivative spectrum value; and fifthly, applying the model to imaging hyperspectral images to obtain an aphid grade distribution map of a field scale.
Further, screening a plurality of sensitive spectral bands of the canopy of the aphid-damaged cotton by using a ratio derivative method.
The application of the method for extracting the spectral characteristics of the cotton canopy of the cotton aphid damage in the bud period and estimating the grade of the cotton canopy is used for assisting accurate and quantitative pesticide application of the cotton aphid damage in the bud period.
The invention has the following beneficial effects:
according to the method for extracting the spectral characteristics and estimating the grade of the cotton canopy of the aphid pests at the bud period of the cotton, the sensitive spectral band of the cotton canopy of the aphid pests at the bud period is extracted, and an aphid grade estimation model based on the sensitive spectrum is constructed, so that the method is used for rapidly monitoring the grade distribution of the aphid pests at the field scale, is convenient for pesticide application and monitoring of the cotton aphid pests, can assist in accurate and quantitative pesticide application, reduces environmental pollution, and has very important significance on actual production.
Drawings
FIG. 1 is a diagram showing the distribution of aphid damage levels during the bud stage of cotton according to the present invention.
FIG. 2 is a formula summary chart of the present invention.
FIG. 3 is a spectrum chart of canopy of different aphid damage grades according to the present invention.
FIG. 4 is a spectrum of the ratio of different grades of stressed cotton plants to healthy plants of the present invention (healthy cotton plant spectra are denominator).
FIG. 5 is a spectrum of the ratio of stressed cotton plants to healthy plants of different grade according to the present invention (spectrum of very severe aphid grade cotton plants is denominator).
FIG. 6 is a graph of the ratio derivative of various aphid damage ratings according to the present invention.
FIG. 7 is a graph showing the correlation between the aphid grade and the ratio derivative spectrum value according to the present invention.
FIG. 8 is a diagram of a first PLSR-based aphid rating estimation model of the present invention (left is spectral reflectance based on three sensitive bands, right is derivative spectral value based on three sensitive band ratio).
FIG. 9 is a diagram of a second PLSR-based aphid rating estimation model according to the present invention (left is a spectrum reflectance based on three sensitive bands, and right is a derivative spectrum value based on three sensitive band ratios).
FIG. 10 is a table of the cotton aphid damage rating prediction model of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-10, the present embodiment provides a method for extracting spectral characteristics and estimating a grade of a canopy of cotton of aphid pests in a bud period of cotton, including the following steps: firstly, acquiring a hyperspectral image of a cotton canopy of aphid damage in a cotton bud period by using a low-altitude unmanned aerial vehicle carried imaging high-spectrum spectrometer; secondly, analyzing canopy spectral response differences of cotton stressed by different aphid levels; thirdly, screening out sensitive spectrum bands of the cotton canopy of the cotton aphid damage in the bud period by using a ratio derivative method according to the hyperspectral image of the cotton canopy of the cotton aphid damage in the bud period; fourthly, constructing an aphid level estimation model based on a partial least square method of the sensitive spectrum ratio derivative spectrum value; and fifthly, applying the model to imaging hyperspectral images to obtain an aphid grade distribution map of a field scale.
The application of the method for extracting the spectral characteristics of the cotton canopy of the cotton aphid damage in the bud period and estimating the grade of the cotton canopy is used for assisting accurate and quantitative pesticide application of the cotton aphid damage in the bud period.
This example was carried out based on the Korla test station (41 ° 44'59"N, 85 ° 48'30" E) implemented in the Kurla region in the middle of Xinjiang and the plant protection institute of the national academy of agriculture of the United states of Lei. The region belongs to temperate continental drought climate, the total sunshine duration of the whole year is 2990 hours, the average annual temperature is 11.4 ℃, the minimum temperature is-28 ℃, the average annual precipitation is 58.6 mm, and the maximum annual evaporation is 2788.2 mm. The cotton is a main local planting crop, the scale is large, the planting structure is simple, and the cotton aphids are main cotton pests in the area. Data collection is carried out in the bud period of cotton in the last ten days of 6 months, a test cell with serious field aphids is selected, and the test area is 100m long and 30m wide. During the test period, no pesticide for inhibiting the growth of aphids is applied, 20 ridges are selected as a data acquisition area (4 corners are positioned by using a GPS) in the middle of a test cell, the length of the area is 100m, the width of the area is 30m, and peripheral cotton is used as a protection row. The cotton to be tested is a test variety of the cotton pest group in the plant protection institute of the agricultural academy of sciences, and is sowed in the middle and last ten days of 4 months, dibbled on a film and drip-irrigated under the film. Cotton aphid-stressed cotton behavior is naturally occurring in the field.
Acquiring a hyperspectral image of a cotton canopy of aphid damage in a cotton bud period by using a low-altitude unmanned aerial vehicle carried imaging high-spectrum spectrometer:
the cotton bud period is positive in the last 6 th month, which is the aphid full-blown period. Because the data of the cotton plant canopy acquired by the unmanned aerial vehicle remote sensing platform are acquired, when the disease grade is investigated and sampled in the field, all the selected sampling points are positioned on the cotton plant top canopy. The grading standard of the degree of damage of cotton aphids refers to the national standard (GB/T15799-2011) and is shown in Table 1. After being infected by aphids. According to the aphid grading standard, a total of 76 sample points are selected, wherein 16 healthy plants are selected, 15 sample points are selected from each level with the aphid grade of 1-4, and the 76 sample points are uniformly distributed in a cotton planting area.
TABLE 1 grading Standard of Aphis gossypii Glover as harmful
| Grade of aphid damage | Standard of |
| 0 | No aphid and |
| 1 | The aphids exist and the leaves are not damaged |
| 2 | The most seriously damaged leaf with aphid has a shriveling or slightly curled, nearly semicircular |
| 3 | There is aphid, the receivingThe most harmful blade is curled into an arc shape up to or above a semicircle |
| 4 | The heaviest leaves with aphids are completely curled and spherical |
This embodiment unmanned aerial vehicle remote sensing platform contains: eight rotor unmanned aerial vehicle, flight control system, formation of image hyperspectral appearance, miniature single-board computer PokiniZ, wireless remote sensing system, ground control system and data processing system. The eight-rotor unmanned aerial vehicle is AZUP-T8 produced by Tiantu company, the length of a single arm is 60cm, the net weight of the unmanned aerial vehicle body is 7kg, the loadable weight is 10kg, and the endurance time is 30 min; the imaging high-resolution spectrometer is CubertUHD185, the shooting mode is full-frame, non-scanning and real-time imaging, the spectral range is 450-950nm, the wave band number is 125, the spectral sampling interval is 4nm, and the spectral resolution is 8 nm. The unmanned aerial vehicle remote sensing operation is clear and cloudless on the day, the wind speed is low, the navigation speed is 6m/s, the navigation height is 50m, the course overlapping degree is 80%, the lateral overlapping degree is 60%, the ground control station remotely controls PokiniZ operation through a wireless network and stores the shooting data in PokiniZ.
UHD185 is a new type of snapshot hyperspectral sensor featuring short exposure acquisition time. The sensor measures 0.47 kg and has dimensions of 195 × 67 × 60 mm. It can obtain radiation reflection from visible light to near infrared spectrum. The sensor can capture wavelengths of 450-950nm while maintaining reasonable image balance and good spectral resolution (4 nm). The radiation was recorded and processed into a 1000 x 1000 (1 band) full color image and a 50 x 50 (125 band) hyperspectral image. The obtained full-color image has rich texture information and is relatively simple to splice; however, the images lack spectral information. A 50 x 50 (125 band) hyperspectral image is characterized by abundant spectral information but lacks texture information. The method comprises the steps of performing hyperspectral image fusion by using Cube-Pilot software developed by Germany Cubert company and Agisosoft PhotoSacan software developed by Agisosoft LLC company, and splicing all images together after fusion.
After stitching the UHD185 images together, an imaged hyperspectral image of ground resolution of 1cm is obtained. Calculating the average coronal layer spectral reflectance of the ROI corresponding to each sample point in the wave band of 450nm-950nm by using an ENVI ROI (ITT Visual Information Solutions, Boulder, CO, USA) tool, and processing to obtain the hyperspectral reflectance data corresponding to 76 sample points.
Insect pest spectral feature extraction based on a ratio derivative method, which is a special spectral processing method provided based on a linear mixed model, can remove background end member influence to obtain continuous spectrum section ratio derivative spectrum, and is an effective method for selecting characteristic wave bands. The method comprises the steps of firstly carrying out band-by-band ratio operation on two continuous spectrums to obtain a ratio spectrum of the two continuous spectrums, and then carrying out derivation operation on the ratio spectrum to obtain a ratio derivative spectrum curve of the two continuous spectrums. The curve can eliminate the influence of other substances in the mixture, directly obtain the corresponding relation between the target object and the change of the mixed spectrum, and extract the wave band which is sensitive to the target information. FIG. 2 formula (1) is a linear spectral mixture model containing m components.
In formula (1) in fig. 2, i =1, 2, …, n is a spectrum band, j =1, 2, 3, …, m is an end member component, and Fj is a proportion of each end member in the mixed pixel. In the case of not considering the error term, when each pixel only contains two substances, the linear spectrum mixed model can be simplified into formula (2) in fig. 2;
when both sides of equation (2) of fig. 2 are simultaneously divided by the spectrum of the 2 nd substance, the equation becomes equation (3) of fig. 2;
the two pairs of the formula (3) in FIG. 2 are derived, and the formula (4) in FIG. 2 is obtained;
it can be seen from equation (4) in fig. 2 that the derivative spectrum is already independent of the proportion of the 2 nd substance, i.e. the value of the spectrum after derivation is only linearly dependent on the proportion of the 1 st substance and not on the proportion of the substance as a divisor.
In this embodiment, each pixel in the imaging spectral data of the unmanned aerial vehicle may be regarded as a mixed pixel of a background plant and a canopy spectrum of a pest-stressed plant, where the background plant is a standard plant not affected by pests, i.e., a relatively healthy plant in a plot. In the embodiment, aphids are the most important stresses, in order to remove the interference of background information, the spectrum of a healthy plant is selected as a background end member, and the spectrum is used as a denominator to perform ratio derivative spectrum processing on imaging spectrum data pixel by pixel, so that the influence of non-pest stress background factors is removed, and the spectrum characteristic only reflecting the pest stress degree of the plant is obtained.
Constructing a model and evaluating the precision; in previous studies, Partial Least Squares Regression (PLSR) method has been widely applied to estimation of vegetation physiological parameters. The partial least squares regression method combines the ideas of Principal Component Analysis (PCA) and Canonical Correlation Analysis (CCA) on the basis of a multiple linear regression method, so that independent variables are independent from each other and the difference between the reaction and the dependent variable can be maximized. Solves the problems of small samples and multiple collinearity which are often encountered in multiple linear regression analysis. Partial least squares regression is expressed in the formulas (5) and (6) in fig. 2;
in formula (5) and (6) in FIG. 2, yiThe target variable is the aphid damage grade (dependent variable); x is the number ofijIs a spectral reflectance value or a spectral ratio derivative value (independent variable) of the band; m is the number of spectral bands; n is the number of samples; e.g. of the typeiIs an error; beta is akIs a regression coefficient TjkIs a latent variable; r is the number of the latent variables; ckjAre latent variable coefficients. In the embodiment, the spectral reflectivity and the ratio derivative spectral value of the sensitive waveband are respectively used as dependent variables, a PLSR prediction model is established, and the cotton aphid damage grade is predicted.
In this embodiment, a single variable inversion model and a multivariate inversion model are constructed by using effective characteristic bands extracted by the reflectance spectrum and the ratio derivative spectrum, respectively.
Analyzing the model precision; coefficient of determination (R), and evaluation criteria for the correlation of the model established by the respective aphid sample level values and the spectral reflectance or spectral ratio derivative values of the sensitive waveband. The calculation formula is shown as formula (7) in FIG. 2;
wherein xi and yi are sample values, n is the number of samples, and x and' y are average values of the samples. The larger R is, the better the correlation among variables participating in evaluation is, and the larger the reference value is; the smaller R is, the poorer correlation between the R and the R is, and the lower reference value is.
Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) are used as criteria for evaluating the estimated and measured disease index values. The calculation formulas are shown in the formulas (8) and (9) of FIG. 2:
where xm represents the analog value, xs represents the measured value, and n is the number of samples. In this embodiment, for the disease index DI, the RMSE is very sensitive to large errors, and a smaller RMSE indicates a higher accuracy of the estimate of DI. The MAE calculates the absolute value of the absolute error of the measured value, and then calculates the average value, the average absolute error is made absolute value due to the deviation, the average error can better reflect the actual situation of the error of the predicted value, and the RMSE and the MAE are used to help better analyze the error and avoid the RMSE from being too large due to individual error.
Analyzing canopy spectral response differences of cotton stressed by different aphid damage grades;
spectral curve characteristics of canopy of different aphid damage grades: the cotton aphid threatens cotton plants, on one hand, a sucking mouth device is inserted into the back or tender head of cotton leaves to suck juice, so that the activity, chlorophyll and water content of the leaves are changed, and the damaged leaves are curled towards the back; on the other hand, honeydew excreted by aphids on the leaf surface breeds mould and influences photosynthesis. In order to clarify the spectrum change of cotton damaged by Aphis gossypii, the present example analyzes and compares the spectra of canopy of cotton with different aphid ratings, as shown in FIG. 3.
In a Visible (VIS) wave band (450 nm-670 nm), under the absorption action of chlorophyll, the canopy spectral reflectance of healthy cotton is lower, and gradually increases with the increase of the aphid damage level of cotton, because the cotton leaves are infected by cotton aphids, the pigment, water and activity of the cotton leaves are reduced, so that the spectral reflectance in a Visible light region is increased; in a red edge A area (670 nm-730 nm), the red edge A area moves to a short wave direction along with the increase of the aphid damage grade, namely 'blue shift'; in a Near Infrared (NIR) wave band (760 nm-950 nm), the healthy plants are controlled by the internal tissue structure of leaves, the water content and the canopy structure, the reflectivity is high, the canopy spectrum reflectivity is gradually reduced along with the increase of the aphid level, the cotton aphid damages the leaf tissue structure, the water content of the leaves is reduced, the leaves are curled, the reflectivity of the near infrared wave band is reduced, the aphid likes a dry environment, the stressed plants are relatively short and sparse, and the sparse space is poor in tightness, so that the aphid propagation and survival are facilitated. Therefore, the higher the aphid grade is, the higher the aphid density is, the more serious the damage is, and the lower the reflectivity of the near-infrared band is.
Based on the spectral feature extraction of the ratio derivative method, the spectrum of each spectrum is respectively treated by taking the spectrum of a healthy cotton plant and the spectrum of a cotton plant stressed by the highest level aphid as divisors according to the formula (3) in fig. 2 to obtain a ratio spectrogram (as shown in fig. 4 and 5). When the spectrum of a healthy cotton plant is taken as a divisor, the strong spectrum characteristic of the stressed cotton plant is highlighted (figure 4)), and it can be seen that the higher the aphid grade is, the more prominent the spectrum characteristic is, the lower the aphid grade is, the gentler the spectrum characteristic is, and the spectrum of the healthy plant tends to be; otherwise, the strong spectrum characteristic of healthy cotton plants is highlighted (fig. 5). In summary, the spectral ratio processing can suppress the spectral features of the components as divisors as background while highlighting the effect of other components on the mixed spectrum.
The spectra in fig. 4 were respectively derived according to equation (4) in fig. 2, to obtain the ratio derivative spectra shown in fig. 4. As mentioned before, the spectrum after derivation is only linearly dependent on the ratio of one substance, and not on the substance that is the divisor. That is, the influence of the background end member substance can be eliminated by processing the mixed spectrum by the ratio derivative method, so that the spectrum value is linearly related to the target substance. FIG. 4 is a ratio derivative spectrum derived from a spectrum obtained by dividing the spectrum of a healthy cotton plant, wherein the ratio derivative spectrum curve is irrelevant to healthy vegetation information and only the stressed vegetation information is left. As can be seen from the figure, the ratio spectral curves of different aphid ratings are all in the "three-edge" region commonly used for characterizing the spectral features of vegetation: three wave crests appear in the areas of the blue edge B, the yellow edge C and the red edge A, and a large number of tests and researches show that three edges representing the spectral characteristic wave band positions of the vegetation play a great role in the research of diagnosing the vegetation suffering from diseases, insect pests, heavy metal pollution and the like. Along with higher aphid grade and stronger stress information, the absolute value of the ratio reciprocal spectrum is higher, and the peak values respectively appear in the wave bands of 514nm, 566nm and 698nm, as shown in the table (2):
TABLE (2) list of aphid damage grade sensitive wave band extracted based on ratio derivative method
| Three-sided region | Wave band range (nm) | Peak wavelength (nm) |
| Blue edge B | 490-530 | 514 |
| Yellow edge C | 550-580 | 566 |
| Red edge A | 670-730 | 698 |
Three sensitive wave bands extracted by using a ratio derivative method respectively appear in the areas of a blue edge B, a yellow edge C and a red edge A, and are consistent with the areas. In order to further verify the three sensitive bands screened by the ratio derivative method, correlation analysis is carried out on the ratio derivative spectral values of the aphid grades of 76 sample points and the 450-fold 950 band, the coefficient correlation diagram is shown in fig. 6, the ratio derivative spectral values of all bands of most bands, yellow-edge C regions and red-edge a regions of the blue-edge B region and the correlation coefficients of the aphid grades all achieve extremely significant correlation relations and are sensitive band regions of aphids, wherein the 514nm, 566nm and 698nm bands have strong correlation in the three regions, and the correlation is consistent with the conclusion.
And (3) establishing an aphid level estimation model and analyzing precision, dividing 76 sampling points into a building module and a verification module on the basis of integrating the analysis results, wherein 51 samples are used for establishing the model, and 25 verification samples are used for establishing the aphid level estimation model. The model construction method comprises the following steps:
(1) the method is characterized in that the level of aphids suffered by cotton is taken as a dependent variable, the reflectivities (marked as R514, R566 and R698) of three sensitive wave bands and the ratio derivative spectral values (marked as DR514, DR566 and DR 698) are taken as independent variables, a univariate linear regression model is constructed, and researches show that the level of aphids suffered by cotton and the levels of R514, R566, R698, DR514, DR566 and DR698 have extremely obvious linear regression relations, and the graph is shown in figure 10. The aphid damage grade estimation model constructed based on the three sensitive waveband reflectivities achieves extremely obvious correlation. Wherein the linear regression model y =34.188x-1.9447 constructed by using the 514 wave band determines the coefficient R =0.4156, and the estimation accuracy is highest; the aphid damage level estimation models constructed based on the three sensitive waveband ratio derivative spectral values also achieve extremely significant correlation, wherein a linear regression model y = -90.734x +0.8942 constructed by utilizing 698 waveband ratio derivative spectral values is adopted, R =0.6005, and the estimation precision is highest.
(2) In order to investigate the estimation accuracy of the multivariate linear regression model on the aphid level of cotton, the present embodiment uses a Partial Least Squares Regression (PLSR) method to select the same modeling and verification samples as a single sensitive band, and constructs an aphid level estimation model by using the reflectance and ratio derivative spectral values of the three sensitive bands as independent variables, so as to obtain a partial minimum regression multiplication model as shown in fig. 8. The best model for estimating the aphid level is the PLSR model Y =0.844858569-34.20159951 × DR514+141.4817237 × DR566-31.66947929 × DR698 constructed by using the ratio derivative spectrum values of the three sensitive bands, and R is 0.6117, and the correlation is superior to not only the PLSR model Y = -0.336589121+93.10261285 × R514-79.35974092 × R566+24.61404936 × R698 constructed by the spectral reflectances of the three sensitive bands, but also the linear regression model constructed by the reflectances of the three sensitive bands and the ratio derivative spectrum values. Fitting analysis is performed by using measured values of 25 test samples and estimated values of two models, and using decision coefficients R, Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) of a fitting equation to examine model estimation capability and precision, a higher R indicates a better fitting effect, and a smaller RMSE and MAE indicates a higher model precision. The test results are shown in fig. 9.
In summary, the accuracy of the univariate model and the multivariate model constructed based on the spectral ratio derivative values is higher than that of the regression model constructed based on the spectral reflectivity. The model constructed by the ratio derivative spectrum values of the multiple variables is higher than the estimation model constructed by the ratio derivative spectrum values of the single wave band as variables. Therefore, the PLSR model constructed based on the ratio derivative spectral values is an optimal model. The aphid level estimation model is applied to the hyperspectral image of the unmanned aerial vehicle to obtain the aphid level distribution situation in the bud period of the cotton in the test area, referring to fig. 1, the insect pest level in the area with darker color is lighter, the insect pest level in the area with lighter color is heavier, and the white area is a bare area.
In a word, the invention obtains the imaging hyperspectral images of the cotton canopy obtained by the low-altitude unmanned machine at different aphid grades in the cotton bud period, namely the full-fleshed aphid period. The method comprises the steps of extracting canopy hyperspectral data of different aphid grades by combining ground survey data, firstly analyzing the cotton canopy spectral characteristics of the different aphid grades, then screening the spectral bands sensitive to aphids by using a ratio derivative method, and finally constructing an aphid grade estimation model by using the ratio derivative values of the sensitive bands.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; 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 (3)
1. The method for extracting the spectral characteristics of the cotton canopy of the aphid pests in the bud period of the cotton and estimating the grade is characterized by comprising the following steps of: firstly, acquiring a hyperspectral image of a cotton canopy of aphid damage in a cotton bud period by using a low-altitude unmanned aerial vehicle carried imaging high-spectrum spectrometer; secondly, analyzing canopy spectral response differences of cotton stressed by different aphid levels; thirdly, screening out sensitive spectrum bands of the cotton canopy of the cotton aphid damage in the bud period by using a ratio derivative method according to the hyperspectral image of the cotton canopy of the cotton aphid damage in the bud period; fourthly, constructing an aphid level estimation model based on a partial least square method of the sensitive spectrum ratio derivative spectrum value; and fifthly, applying the model to imaging hyperspectral images to obtain an aphid grade distribution map of a field scale.
2. The method for extracting spectral features and estimating the grade of the canopy of cotton damaged by aphids at the bud stage of the cotton as claimed in claim 1, wherein: screening a plurality of sensitive spectral bands of the canopy of the aphid-damaged cotton by adopting a ratio derivative method.
3. The use of the method for extracting spectral features and estimating the grade of the canopy of cotton aphid damage at the bud stage of cotton as claimed in claim 1, is characterized in that: is used for auxiliary accurate and quantitative pesticide application of cotton bud aphid pests.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911087521.4A CN110658211A (en) | 2019-11-08 | 2019-11-08 | Method for extracting spectral characteristics of cotton canopy of aphid pests in cotton bud period and estimating grade |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911087521.4A CN110658211A (en) | 2019-11-08 | 2019-11-08 | Method for extracting spectral characteristics of cotton canopy of aphid pests in cotton bud period and estimating grade |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110658211A true CN110658211A (en) | 2020-01-07 |
Family
ID=69043659
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911087521.4A Pending CN110658211A (en) | 2019-11-08 | 2019-11-08 | Method for extracting spectral characteristics of cotton canopy of aphid pests in cotton bud period and estimating grade |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110658211A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111062251A (en) * | 2020-03-23 | 2020-04-24 | 乔红波 | Monitoring method of farmland cotton aphid pest grade model based on unmanned aerial vehicle imaging |
| CN112257690A (en) * | 2020-12-21 | 2021-01-22 | 航天宏图信息技术股份有限公司 | Heavy metal pollution assessment method and device |
| CN112528726A (en) * | 2020-10-14 | 2021-03-19 | 石河子大学 | Aphis gossypii insect pest monitoring method and system based on spectral imaging and deep learning |
| CN112782103A (en) * | 2021-02-01 | 2021-05-11 | 石河子大学 | Method and system for monitoring early damage of cotton aphids on leaf blades in seedling stage of cotton |
| CN114694020A (en) * | 2022-03-04 | 2022-07-01 | 新疆农业大学 | Construction method of cotton aphid remote sensing forecast model |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103278460A (en) * | 2013-05-30 | 2013-09-04 | 华南农业大学 | Test and analysis method of red spider insect pest coercion conditions of orange trees |
| CN104266982A (en) * | 2014-09-04 | 2015-01-07 | 浙江托普仪器有限公司 | Large-area insect pest quantization monitoring system |
| CN106290197A (en) * | 2016-09-06 | 2017-01-04 | 西北农林科技大学 | The estimation of rice leaf total nitrogen content EO-1 hyperion and estimation models construction method |
| CN107024439A (en) * | 2017-03-23 | 2017-08-08 | 西北农林科技大学 | A kind of paddy rice different growing chlorophyll content EO-1 hyperion estimating and measuring method |
| CN107704835A (en) * | 2017-10-16 | 2018-02-16 | 北京市遥感信息研究所 | A kind of method using spectral remote sensing image recognition sea Artificial facilities |
| CN108375550A (en) * | 2018-01-12 | 2018-08-07 | 河南农业大学 | The construction method of winter wheat full rot disease Index Prediction Model based on spectral index and application |
| CN108694391A (en) * | 2018-05-16 | 2018-10-23 | 黄铁成 | Populus Euphratica spring looper disaster monitoring method based on high-spectrum remote-sensing |
| CN110361344A (en) * | 2019-08-30 | 2019-10-22 | 北京麦飞科技有限公司 | Degree of disease diagnostic method based on EO-1 hyperion |
-
2019
- 2019-11-08 CN CN201911087521.4A patent/CN110658211A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103278460A (en) * | 2013-05-30 | 2013-09-04 | 华南农业大学 | Test and analysis method of red spider insect pest coercion conditions of orange trees |
| CN104266982A (en) * | 2014-09-04 | 2015-01-07 | 浙江托普仪器有限公司 | Large-area insect pest quantization monitoring system |
| CN106290197A (en) * | 2016-09-06 | 2017-01-04 | 西北农林科技大学 | The estimation of rice leaf total nitrogen content EO-1 hyperion and estimation models construction method |
| CN107024439A (en) * | 2017-03-23 | 2017-08-08 | 西北农林科技大学 | A kind of paddy rice different growing chlorophyll content EO-1 hyperion estimating and measuring method |
| CN107704835A (en) * | 2017-10-16 | 2018-02-16 | 北京市遥感信息研究所 | A kind of method using spectral remote sensing image recognition sea Artificial facilities |
| CN108375550A (en) * | 2018-01-12 | 2018-08-07 | 河南农业大学 | The construction method of winter wheat full rot disease Index Prediction Model based on spectral index and application |
| CN108694391A (en) * | 2018-05-16 | 2018-10-23 | 黄铁成 | Populus Euphratica spring looper disaster monitoring method based on high-spectrum remote-sensing |
| CN110361344A (en) * | 2019-08-30 | 2019-10-22 | 北京麦飞科技有限公司 | Degree of disease diagnostic method based on EO-1 hyperion |
Non-Patent Citations (11)
| Title |
|---|
| LIU ZHAN-YU等: "Characterizing and Estimating Fungal Disease Severity of Rice Brown Spot with Hyperspectral Reflectance Data", 《RICE SCIENCE》 * |
| SU JINYA等: "Wheat yellow rust monitoring by learning from multispectral UAV aerial imagery", 《COMPUTERS AND ELECTRONICS IN AGRICULTURE》 * |
| 乔红波等: "基于支持向量机模型的冬小麦全蚀病为害等级遥感监测", 《麦类作物学报》 * |
| 甘甫平等: "《遥感岩矿信息提取基础与技术方法研究》", 29 February 2004, 地质出版社 * |
| 程登发等: "《农作物重大生物灾害监测与预警技术》", 31 December 2014, 重庆出版社 * |
| 程起敏: "《遥感图像检索技术》", 31 May 2011, 武汉大学出版社 * |
| 罗红霞等: "基于高光谱遥感技术的农作物病虫害应用研究现状", 《广东农业科学》 * |
| 赵恒谦: "《高光谱矿物定量反演模型及不确定性研究》", 31 January 2019, 测绘出版社 * |
| 赵恒谦等: "基于比值导数光谱法的强线性波段特征分析", 《红外与毫米波学报》 * |
| 郭伟等: "基于无人机高光谱影像的冬小麦全蚀病监测模型研究", 《农业机械学报》 * |
| 郭泺等: "《环境空间信息技术原理与应用》", 30 September 2011, 中国环境科学出版社 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111062251A (en) * | 2020-03-23 | 2020-04-24 | 乔红波 | Monitoring method of farmland cotton aphid pest grade model based on unmanned aerial vehicle imaging |
| CN112528726A (en) * | 2020-10-14 | 2021-03-19 | 石河子大学 | Aphis gossypii insect pest monitoring method and system based on spectral imaging and deep learning |
| CN112257690A (en) * | 2020-12-21 | 2021-01-22 | 航天宏图信息技术股份有限公司 | Heavy metal pollution assessment method and device |
| CN112782103A (en) * | 2021-02-01 | 2021-05-11 | 石河子大学 | Method and system for monitoring early damage of cotton aphids on leaf blades in seedling stage of cotton |
| CN112782103B (en) * | 2021-02-01 | 2023-12-01 | 石河子大学 | Method and system for monitoring early damage of cotton aphids on cotton seedling stage leaves |
| CN114694020A (en) * | 2022-03-04 | 2022-07-01 | 新疆农业大学 | Construction method of cotton aphid remote sensing forecast model |
| CN114694020B (en) * | 2022-03-04 | 2024-04-12 | 新疆农业大学 | Construction method of cotton aphid remote sensing prediction model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110658211A (en) | Method for extracting spectral characteristics of cotton canopy of aphid pests in cotton bud period and estimating grade | |
| CN107796764B (en) | Method for constructing wheat leaf area index estimation model based on three-band vegetation index | |
| Atzberger et al. | Comparative analysis of three chemometric techniques for the spectroradiometric assessment of canopy chlorophyll content in winter wheat | |
| CN110376167A (en) | Rice leaf nitrogen content monitoring method based on unmanned plane EO-1 hyperion | |
| CN111091052A (en) | Corn lodging area extraction system and method based on maximum likelihood method | |
| Wiegand et al. | Spectral observations for estimating the growth and yield of rice | |
| CN111062251A (en) | Monitoring method of farmland cotton aphid pest grade model based on unmanned aerial vehicle imaging | |
| Matese et al. | Mapping of vine vigor by UAV and anthocyanin content by a non-destructive fluorescence technique | |
| CN113268923A (en) | Summer corn yield estimation method based on simulation multispectral | |
| Warren et al. | Agricultural applications of high-resolution digital multispectral imagery | |
| CN116645603A (en) | Soybean planting area identification and area measurement method | |
| CN114140695B (en) | Prediction method and system for tea tree nitrogen diagnosis and quality index determination based on unmanned aerial vehicle multispectral remote sensing | |
| Tavakoli et al. | Evaluation of different sensing approaches concerning to nondestructive estimation of leaf area index (LAI) for winter wheat | |
| CN111275567A (en) | SENP cotton yield estimation method based on unmanned aerial vehicle image and estimation model construction method | |
| CN112861712B (en) | Agricultural pest and disease monitoring method based on artificial intelligence and multi-temporal remote sensing | |
| Ranjan et al. | Thermal-RGB imagery and in-field weather sensing derived sweet cherry wetness prediction model | |
| CN112345467B (en) | Model for estimating physiological parameters of rice by using remote sensing technology and application thereof | |
| CN114460015A (en) | Method and model for predicting chlorophyll content of canopy of rice in cold region | |
| CN114694020B (en) | Construction method of cotton aphid remote sensing prediction model | |
| CN111537510A (en) | Farmland protection forest wind-prevention effect metering method based on spatial information technology | |
| CN117011694A (en) | Forest tree growth parameter prediction method based on cascade circulation network | |
| Liu et al. | Estimating wheat coverage using multispectral images collected by unmanned aerial vehicles and a new sensor | |
| Li et al. | Optimizing spectral index to estimate the relative chlorophyll content of the forest under the damage of Erannis jacobsoni Djak in Mongolia | |
| CN114694041A (en) | Hyperspectral identification method for cotton phytotoxicity and spider mite insect damage | |
| CN115761542A (en) | Winter wheat ear stage pest and disease degree evaluation method based on unmanned aerial vehicle remote sensing image |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200107 |