Fertilization prescription map rapid generation method and system based on soil spectrum detection

Abstract

The invention provides a method and a system for rapidly generating a fertilization prescription map based on soil spectrum detection. The method comprises the following specific steps: S1, acquiring data required by the method; s2, data preprocessing; s3, soil spectrum characteristic wave band selection; s4, establishing a soil rapid detection model based on a frost ice optimization algorithm; s5, generating a soil nutrient distribution map; s6, establishing a fertilization model; the method has the advantages that the soil data of the field is acquired, the soil nutrient detection model is established, the soil nutrient distribution map is generated, and the soil variable fertilization prescription map is generated according to the target yield and the soil nutrient content; the prescription map generation speed is increased from the soil detection link by establishing the soil nutrient rapid detection model, the variable fertilization prescription map generation method is simplified, the fertilization decision-making efficiency is improved, scientific and reasonable fertilizer application is achieved, and the fertilizer utilization rate is increased.

Description

Technical Field

[0001] The present invention relates to the technical field of variable fertilization, and in particular to a method and system for quickly generating a fertilization prescription map based on soil spectrum detection. Background Art

[0002] Fertilizer is an important material for agricultural production. Accurate and reasonable variable fertilization is an effective means to increase crop yields and fertilizer utilization. Today, my country's fertilizer utilization is much lower than that of other countries. For this reason, my country has conducted research on precision agriculture, and variable fertilization is used as the main technical means for fertilization. Variable fertilization technology based on prescription maps and real-time variable fertilization technology based on conductivity are used in variable fertilization. Variable fertilization based on prescription maps can scientifically fertilize by understanding the nutrient distribution of the field in advance, which reduces the hardware requirements of the equipment and is more suitable for promotion and implementation.

[0003] However, the prescription generation process of variable fertilization technology based on prescription maps requires soil sampling and soil nutrient chemical analysis, which consumes a lot of manpower, material resources and time costs, and it is easy to miss the best time for fertilization. Summary of the invention

[0004] To this end, an object of the present invention is to propose a method and system for quickly generating a fertilization prescription map based on soil spectral detection, so as to solve the problems mentioned in the background technology and overcome the shortcomings of the prior art.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present invention provides a method for quickly generating a fertilization prescription map based on soil spectrum detection, comprising:

[0007] The second spectral data of the soil to be tested and the first soil data are subjected to feature selection through a competitive adaptive reweighted sampling algorithm to screen out characteristic bands; the first soil data includes nitrogen, phosphorus and potassium nutrient content data of the soil to be tested;

[0008] The characteristic bands were used as input and the soil nitrogen, phosphorus and potassium nutrient content was used as output. The optimal parameters of convolution-support vector regression were screened out through the frost optimization algorithm to establish a regression model.

[0009] Obtain soil nutrients based on the regression model and generate a soil nutrient distribution map;

[0010] Establish a target yield fertilization model;

[0011] Divide the soil nutrient distribution map into grids, obtain the required fertilizer amount for a single grid through the target yield fertilization model, and generate a complete fertilization prescription map;

[0012] The target yield fertilization model includes: a nitrogen application amount model required for the target yield, a phosphorus application amount model required for the target yield, and a potassium application amount model required for the target yield;

[0013] The model of nitrogen application amount required for target yield is as follows:

[0014]

[0015] N s =N t1 ×K×2.25

[0016]

[0017] Among them, U N is the amount of nitrogen required for the target yield, u is the target yield, N r The amount of nitrogen required per 100 kg of crops, N s The nitrogen supply of soil nutrients, N t1 is the average soil nitrogen content detection value, K is the soil correction coefficient, W N is the nitrogen application rate, C N is the effective content of nitrogen fertilizer in fertilizer, R N is the nitrogen fertilizer utilization efficiency.

[0018] Preferably, the method further includes data acquisition, which includes: acquiring soil to be tested, air-drying, grinding, and sieving the soil to be tested, and acquiring first spectral data and first soil data of the processed soil to be tested.

[0019] Preferably, the method further includes data preprocessing, which includes: converting the acquired processed first spectral data of the soil to be tested into an xlsx file, and performing averaging processing and second-order difference preprocessing operations after the conversion is completed to obtain second spectral data.

[0020] Preferably, the soil is sieved with a 0.25 mm standard sieve, the first spectral data is acquired by an ASD ground feature spectrometer, 3 spectral curves are acquired for a single sample, the sampling interval is 1 nm, and whiteboard correction is performed before each sampling.

[0021] Preferably, the converting of the acquired first spectral data of the processed soil to be tested into an xlsx file comprises: converting the spectral ASD file data into an xlsx file by using ENVI software, wherein the data is stored in a form of: one row is the reflectance of one spectral band, and every three columns are one sample.

[0022] Preferably, the second spectral data and the first soil data are subjected to feature selection through a competitive adaptive reweighted sampling algorithm to screen out characteristic bands, including: running the competitive adaptive reweighted sampling characteristic band screening algorithm through the libpls package of MATLAB, using Monte Carlo cross-validation to determine the number of principal components of the competitive adaptive reweighted sampling, and finally obtaining the best characteristic band.

[0023] Preferably, the model of phosphorus application amount required for the target yield is as follows:

[0024]

[0025] P s =P t1 ×K×2.25

[0026]

[0027] Among them, U P is the phosphorus amount required for the target yield, u is the target yield, P r is the amount of phosphorus required per 100 kg of crops, P s The phosphorus supply of soil nutrients, P t1 is the average soil phosphorus content detection value, K is the soil correction coefficient, W P is the phosphorus application rate, C P is the effective content of phosphorus fertilizer in fertilizer; R P is the utilization rate of phosphorus fertilizer;

[0028] The model of potassium application amount required for target yield is as follows:

[0029]

[0030] K s =K t1 ×K×2.25

[0031]

[0032] Among them, U K is the amount of potassium required for the target yield, u is the target yield, K r The amount of potassium required per 100 kg of crops, K s Potassium supply for soil nutrients, K t1 is the average potassium content in the soil, K is the soil correction factor, and W is the average potassium content in the soil. K is the potassium application rate, C K is the effective content of potassium fertilizer in fertilizer, R K It is the utilization rate of potassium fertilizer.

[0033] Preferably, the method of obtaining soil nutrients according to the regression model and generating a soil nutrient distribution map includes: uniformly obtaining a number of soil samples from multiple soil sampling points in the farmland area by a grid sampling method and recording their longitude and latitude coordinates, obtaining the third spectral data of the soil in the corresponding grid by an ASD ground feature spectrometer, obtaining its characteristic band, and inputting it into the regression model to quickly obtain the nitrogen, phosphorus, and potassium nutrient contents of the soil samples at each grid sample point, putting the longitude and latitude coordinates of each sample point and the nitrogen, phosphorus, and potassium nutrient contents of the soil samples at the sample points into Excel, uploading it to Arcmap software, and generating a soil nutrient distribution map by Kriging spatial interpolation.

[0034] Preferably, the soil nutrient distribution map is divided into grids, and the required fertilizer amount for a single grid is obtained through a target yield fertilization model to generate a complete fertilization prescription map, including: dividing the soil nutrient distribution map into a number of 2m*10m equal-sized rectangular grids, and obtaining the average nitrogen, phosphorus and potassium content of a single grid, traversing all grids, and substituting the average nitrogen, phosphorus and potassium contents of all grids into the target yield fertilization model formula, respectively, to obtain the required fertilizer amount for all grids, thereby completing the generation of a variable fertilization prescription map.

[0035] In a second aspect, the present invention also provides a rapid generation system of a fertilization prescription map based on soil spectrum detection, comprising: a characteristic band screening module, a regression model module, a distribution map generation module, a target yield fertilization model module and a prescription map generation module.

[0036] The characteristic band screening module is configured to screen out characteristic bands by inputting the second spectral data and the first soil data into a competitive adaptive reweighted sampling algorithm for feature selection;

[0037] The regression model module is configured to establish a regression model by taking the characteristic band as input and the soil nitrogen, phosphorus and potassium nutrient content as output, and screening out the best parameters of convolution-support vector regression through the frost and ice optimization algorithm;

[0038] The distribution map generation module is configured to obtain soil nutrients according to the regression model and generate a soil nutrient distribution map;

[0039] The target yield fertilization model module is configured to establish a target yield fertilization model;

[0040] The prescription map generation module is configured to divide the soil nutrient distribution map into grids, obtain the fertilizer amount required for a single grid through the target yield fertilization model, and generate a complete fertilization prescription map;

[0041] The target yield fertilization model includes: a nitrogen application amount model required for the target yield, a phosphorus application amount model required for the target yield, and a potassium application amount model required for the target yield;

[0042] The model of nitrogen application amount required for target yield is as follows:

[0043]

[0044] N s =N t1 ×K×2.25

[0045]

[0046] Among them, U N is the amount of nitrogen required for the target yield, u is the target yield, N r The amount of nitrogen required per 100 kg of crops, N s The nitrogen supply of soil nutrients, N t1 is the average soil nitrogen content detection value, K is the soil correction coefficient, W N is the nitrogen application rate, C N is the effective content of nitrogen fertilizer in fertilizer, R N is the nitrogen fertilizer utilization efficiency.

[0047] In a third aspect, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a method for rapidly generating a fertilization prescription map based on soil spectral detection as described above are implemented.

[0048] In a fourth aspect, the present invention further provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for rapidly generating a fertilization prescription map based on soil spectral detection as described above.

[0049] Therefore, the present invention has the following beneficial effects:

[0050] The present invention obtains characteristic bands between spectra and different soil nutrients, and inputs the reflectance of the characteristic bands into a regression model optimized by a frost optimization algorithm to quickly obtain soil nutrient data, generate a soil nutrient distribution map, and calculate the amount of fertilizer in combination with a corresponding fertilization model, thereby generating a fertilization prescription map, providing fertilization decisions for soil fertilization, and eliminating the need for chemical analysis of the soil in the traditional prescription generation link. This improves the speed of obtaining soil nutrient distribution maps, improves the efficiency of generating fertilization prescription maps, greatly shortens the generation time of fertilization prescription maps, makes it easy to grasp the optimal fertilization time, and makes it easier to promote the implementation of precision agriculture. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flow chart of the method of the present invention;

[0052] Figure 2 is an overall flow chart of the method of an embodiment of the present invention;

[0053] Figure 3is a soil nitrogen distribution diagram of an embodiment of the present invention;

[0054] Figure 4 is a fertilization prescription diagram of an embodiment of the present invention;

[0055] Figure 5 It is a system structure block diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0056] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0057] The terms "first" and "second" in the specification, claims and drawings of this application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.

[0058] like Figure 1 As shown, the present invention provides a method for quickly generating a fertilization prescription map based on soil spectrum detection, comprising:

[0059] The second spectral data and the first soil data are input into a competitive adaptive reweighted sampling algorithm for feature selection to screen out characteristic bands;

[0060] The characteristic bands were used as input and the soil nitrogen, phosphorus and potassium nutrient content was used as output. The optimal parameters of convolution-support vector regression were screened out through the frost optimization algorithm to establish a regression model.

[0061] Obtain soil nutrients based on the regression model and generate a soil nutrient distribution map;

[0062] Establish a target yield fertilization model;

[0063] Divide the soil nutrient distribution map into grids, obtain the required fertilizer amount for a single grid through the target yield fertilization model, and generate a complete fertilization prescription map;

[0064] The target yield fertilization model includes: the target yield required nitrogen application model, the target yield required phosphorus application model and the target yield required potassium application model;

[0065] The model of nitrogen application amount required for target yield is as follows:

[0066]

[0067] N s =N t1 ×K×2.25

[0068]

[0069] Among them, Ux is the nitrogen amount required for the target yield, u is the target yield, N r The amount of nitrogen required per 100 kg of crops, N s The nitrogen supply of soil nutrients, N t1 is the average soil nitrogen content detection value, K is the soil correction coefficient, W N is the nitrogen application rate, C N is the effective content of nitrogen fertilizer in fertilizer, R N is the nitrogen fertilizer utilization efficiency.

[0070] The present invention obtains characteristic bands between spectra and different soil nutrients, and inputs the reflectance of the characteristic bands into a regression model optimized by a frost optimization algorithm to quickly obtain soil nutrient data, generate a soil nutrient distribution map, and calculate the amount of fertilizer in combination with a corresponding fertilization model, thereby generating a fertilization prescription map, providing fertilization decisions for soil fertilization, and eliminating the need for chemical analysis of the soil in the traditional prescription generation link. This improves the speed of obtaining soil nutrient distribution maps, improves the efficiency of generating fertilization prescription maps, greatly shortens the generation time of fertilization prescription maps, makes it easy to grasp the optimal fertilization time, and makes it easier to promote the implementation of precision agriculture.

[0071] Preferably, the method also includes data acquisition, which includes: obtaining the soil to be tested, air-drying, grinding, and sieving the soil to be tested, and obtaining first spectral data and first soil data of the processed soil to be tested, wherein the first soil data includes nitrogen, phosphorus, and potassium nutrient content data of the processed soil to be tested.

[0072] Preferably, the method further includes data preprocessing, which includes: converting the acquired processed first spectral data of the soil to be tested into an xlsx file, and performing averaging processing and second-order difference preprocessing operations after the conversion is completed to obtain second spectral data.

[0073] Preferably, the soil is sieved with a 0.25 mm standard sieve, the first spectral data is acquired by an ASD ground feature spectrometer, three spectral curves are acquired for a single sample, the sampling interval is 1 nm, and whiteboard correction is performed before each sampling.

[0074] Preferably, converting the obtained first spectral data of the processed soil to be tested into an xlsx file includes: converting the spectral ASD file data into an xlsx file through ENVI software, and the data is stored in the form of: one line is the reflectance of one spectral band, and every three columns are one sample.

[0075] Preferably, the second spectral data and the first soil data are input into a competitive adaptive reweighted sampling algorithm for feature selection, and the characteristic bands are screened out, including: running the competitive adaptive reweighted sampling characteristic band screening algorithm through MATLAB's libpls package, using Monte Carlo cross-validation to determine the number of principal components of the competitive adaptive reweighted sampling, and finally obtaining the best characteristic band.

[0076] As a preferred method, the model of phosphorus application required for target yield is as follows:

[0077]

[0078] P s =P t1 ×K×2.25

[0079]

[0080] Among them, U P is the phosphorus amount required for the target yield, u is the target yield, P r is the amount of phosphorus required per 100 kg of crops, P s The phosphorus supply of soil nutrients, P t1 is the average soil phosphorus content detection value, K is the soil correction coefficient, W P is the phosphorus application rate, C P is the effective content of phosphorus fertilizer in fertilizer; R P is the utilization rate of phosphorus fertilizer;

[0081] The model of potassium application amount required for target yield is as follows:

[0082]

[0083] K s =K t1 ×K×2.25

[0084]

[0085] Among them, U K is the amount of potassium required for the target yield, u is the target yield, K r The amount of potassium required per 100 kg of crops, K s Potassium supply for soil nutrients, K t1 is the average potassium content in the soil, K is the soil correction factor, and W is the average potassium content in the soil. K is the potassium application rate, CK is the effective content of potassium fertilizer in fertilizer, R K It is the utilization rate of potassium fertilizer.

[0086] Preferably, obtaining soil nutrients according to the regression model and generating a soil nutrient distribution map includes: uniformly obtaining a number of soil samples from multiple soil sampling points in the farmland area by a grid sampling method and recording their longitude and latitude coordinates, obtaining the third spectral data of the soil in the corresponding grid by an ASD ground feature spectrometer, obtaining its characteristic band, inputting it into the regression model, quickly obtaining the nitrogen, phosphorus and potassium nutrient contents of the soil samples at each grid sample point, putting the longitude and latitude coordinates of each sample point and the nitrogen, phosphorus and potassium nutrient contents of the soil samples at the sample point into Excel, uploading it to Arcmap software, and generating a soil nutrient distribution map by Kriging spatial interpolation.

[0087] Preferably, the soil nutrient distribution map is divided into grids, and the required fertilizer amount for a single grid is obtained through the target yield fertilization model. Generating a complete fertilization prescription map includes: dividing the soil nutrient distribution map into a number of 2m*10m equal-sized rectangular grids, and obtaining the average nitrogen, phosphorus, and potassium content of a single grid, traversing all grids, and substituting the average nitrogen, phosphorus, and potassium contents of all grids into the target yield fertilization model formula to obtain the required fertilizer amount for all grids, thereby completing the generation of a variable fertilization prescription map.

[0088] Next, the method for quickly generating a fertilization prescription map based on soil spectrum detection of the present invention is described in detail in the form of a step flow and in combination with the drawings of the specification.

[0089] like Figure 2 As shown, a method for quickly generating a fertilization prescription map based on soil spectrum detection includes steps S1-S7.

[0090] S1: Acquiring data required by the method, specifically including: acquiring the soil to be tested; air-drying, grinding, and sieving the soil to be tested; acquiring spectral data of the treated soil to be tested, i.e., first spectral data; acquiring nitrogen, phosphorus, and potassium nutrient content information of the treated soil to be tested, i.e., first soil data.

[0091] In one embodiment, specifically, the air-dried soil is ground and sieved through a 0.25 mm standard sieve. The sieved soil is divided into two even parts using a quartering method. One part is used to detect the content of soil nitrogen, phosphorus, and potassium nutrients, which is obtained using a soil nutrient rapid detector; the other part is used to obtain soil spectral data, which is obtained using an ASD ground feature spectrometer. Three spectral curves are obtained for a single sample, and the sampling interval is 1 nm. Whiteboard correction is used before each sampling.

[0092] S2: Data preprocessing, including converting the acquired soil spectral data to an xlsx file, performing averaging processing after the conversion, and performing second-order difference preprocessing operations.

[0093] In one embodiment, specifically, the spectrum asd file is converted into an xlsx file by using ENVI software, and the data is stored in the form of reflectance of one spectral band in one row, and every three columns are one sample. The first column is the spectral band, and the spectral band range is 350-2500nm, through the following Excel formula:

[0094] AVERAGE(OFFSET($B1,,(COLUMNS($B$1:B1)-1)*3,,3)) calculates the average hyperspectral reflectance of each sample. B1 is the cell where the reflectance of the first band of the first sample is located. Quick fill calculation is then used. After the calculation is completed, the data is transposed and converted into a form in which one row is one sample and one column is one spectral band, which is convenient for subsequent data processing and analysis.

[0095] In one embodiment, the converted data is screened to remove the two end bands to reduce the noise impact during acquisition. The removed bands are in the range of 350-390nm and 2451-2500nm. The smoothing operation is performed using the sg_smooth function of MATLAB, and the diff function is used to perform second-order differential derivative on the smoothed data.

[0096] S3: soil spectral feature band selection, including inputting the preprocessed spectral data (i.e., the second spectral data) and the soil nitrogen, phosphorus, and potassium nutrient content data (i.e., the first soil data) into the competitive adaptive reweighted sampling (cars) feature selection algorithm to screen out the feature bands that are most suitable for building the model.

[0097] In one embodiment, the cars characteristic band screening algorithm is run through the libpls package of MATLAB to reduce the dimension of the spectral data, and determine the optimal parameters with the preprocessed spectral reflectance as input and the soil nitrogen, phosphorus and potassium nutrients as output, wherein the number of cars principal components is determined by Monte Carlo cross-validation, and the spectral characteristic bands of nitrogen, phosphorus and potassium are obtained respectively.

[0098] S4: Establish a soil rapid detection model based on the frost and ice optimization algorithm: take the bands after feature selection as input, and the soil nitrogen, phosphorus, and potassium nutrient content data as output, and use the frost and ice optimization algorithm to screen out the optimal parameters of support vector regression (SVR) to establish a regression model, namely the soil rapid detection model.

[0099] In one embodiment, the ratio of training set to test set is 3:1, and the division method adopts stratified sampling to ensure the uniformity of data distribution, and the data set is normalized from 0 to 1. The population size, number of iterations, number of optimization parameters, upper limit of parameter values, and lower limit of parameter values ​​of the frost and ice optimization algorithm are set. The population size is set to the number of features, the number of iterations is 20, and the number of optimization parameters is 3. The parameters of the optimized SVR model are the weight upper limit (BoxConstraint), the feature scaling factor of the kernel function (KernelScale), and the insensitive interval parameter (Epsilon). The weight upper limit range is set to [0.1, 100], the feature scaling factor range of the function is [0.01, 10], and the insensitive interval parameter range is [0.001, 1]. The optimal weight upper limit, feature scaling factor of the kernel function, and insensitive interval parameter of SVR are obtained within the above setting range through the frost and ice optimization algorithm, and are input into the support vector machine for model training. The support vector machine adopts a linear kernel function to complete the establishment of a rapid soil nutrient detection model.

[0100] S5: Generate soil nutrient distribution map: obtain spectral data (i.e., third spectral data) of multiple soil sampling points in the farmland area, input the established regression model, obtain its soil nutrients, and generate a soil nutrient distribution map through a spatial interpolation algorithm based on its geographic coordinates.

[0101] In one embodiment, soil sampling adopts a grid sampling method, and a number of samples are evenly obtained and their geographical location coordinates are recorded. The spectral data is obtained through an ASD ground feature spectrometer. The spectral data is reduced in dimension through a cars feature selection algorithm to obtain its characteristic bands. The nitrogen, phosphorus, and potassium characteristic bands are input into a regression model to quickly obtain the nitrogen, phosphorus, and potassium contents of each sample point. The longitude and latitude coordinates and nutrient content of each sample point are put into Excel and uploaded to Arcmap software. The soil nitrogen content distribution map, soil phosphorus content distribution map, and soil potassium content distribution map are generated through Kriging spatial interpolation.

[0102] In one embodiment, the latitude and longitude coordinates of the sample points and their corresponding nutrient contents are put into Excel and uploaded to Arcmap software. Taking nitrogen nutrient as an example, the nitrogen nutrient field is selected, and the point-to-surface conversion is realized through Kriging spatial interpolation to generate a soil nitrogen content distribution map as shown in FIG. Figure 3 As shown in the figure, the lower left corner is the nutrient classification interval, which evenly divides the nitrogen nutrient content into five levels, and the color from light to dark corresponds to the nitrogen content from low to high. From the figure, we can see the distribution of nitrogen nutrient content in the field, and obtain the soil nutrient nitrogen content values ​​of each area of ​​the whole field to calculate the amount of nitrogen fertilizer required in different areas.

[0103] In one embodiment, the reflectance of the characteristic band of the target area in S3 is obtained by remote sensing, and smoothing-second-order difference preprocessing is performed on it, and input into the regression model to obtain the distribution maps of nitrogen, phosphorus and potassium content in the target area.

[0104] S6: Establish a fertilization model: Establish a target yield fertilization model based on historical farmland yield data, crop types, and fertilizer utilization rates.

[0105] In one embodiment, the target yield fertilization model includes a target yield required nitrogen application amount model, a target yield required phosphorus application amount model and a target yield required potassium application amount model, which are respectively used for calculation.

[0106] In one embodiment:

[0107] Calculate the amount of nitrogen required for target yield (kg / hm 2 ):

[0108] Soil nutrient nitrogen supply: N s =N t1 ×K×2.25;

[0109] The amount of nitrogen required in the current area:

[0110] Among them, U N is the amount of nitrogen required for the target yield, u is the target yield, which is determined based on the average crop yield in the past three years; N r The amount of nitrogen required for every 100 kg of crops can be found by searching according to the crop type; N s The nitrogen supply of soil nutrients, N t1 is the average soil nitrogen content detection value, K is the soil correction coefficient, W N is the nitrogen application rate, C N is the effective nitrogen content in fertilizer, RN 为 Nitrogen fertilizer utilization efficiency.

[0111] In one embodiment, the soil correction coefficient K needs to be prepared in advance by experiments, based on the nitrogen content N0 absorbed by crops in areas where nitrogen fertilizer is not applied: The soil nitrogen content in the unfertilized area is N t0 Get, K = N0 ÷ N t0 ×2.25, to obtain the nitrogen content N0 absorbed by crops in the nitrogen fertilizer application area and the soil nitrogen content N in the unfertilized area in multiple different areas of the farmland t0 , with K as the y-axis, N t0 Draw a scatter plot for the x-axis and perform regression analysis to obtain the regression equation y=ax+b, where the a value is the soil calibration coefficient K.

[0112] In one embodiment, nitrogen fertilizer utilization rate (%), N1 is the nitrogen content absorbed by crops in the nitrogen fertilizer application area: F n The amount of nitrogen fertilizer applied.

[0113] In one embodiment:

[0114] Calculate the phosphorus required for target yield (kg / hm 2 ):

[0115] Soil nutrient phosphorus supply: P s =P t1 ×K×2.25;

[0116] The amount of phosphorus required in the current area:

[0117] Among them, U P is the phosphorus required for the target yield, u is the target yield, which is determined based on the average crop yield in the past three years; P r The amount of phosphorus required for every 100 kg of crops can be found by searching according to the crop type; P s The phosphorus supply of soil nutrients, P t1 is the average soil phosphorus content detection value, K is the soil correction coefficient, W P is the phosphorus application rate, C P is the effective content of phosphorus fertilizer in fertilizer; R P is the utilization rate of phosphorus fertilizer.

[0118] In one embodiment, K is the soil effective nutrient correction coefficient, which needs to be prepared in advance by experiments, based on the phosphorus content P0 absorbed by crops in areas where no phosphorus fertilizer is applied: The soil phosphorus content in the unfertilized area is P t0 Get, K = P0 ÷ P t0 ×2.25, to obtain the phosphorus content P0 absorbed by crops in the phosphate fertilizer application area and the soil phosphorus content detection value P in the unfertilized area in multiple different areas of the farmland t0 , with K as the y-axis, P t0 Draw a scatter plot for the x-axis and perform regression analysis to obtain the regression equation y=ax+b, where the a value is the soil calibration coefficient K.

[0119] In one embodiment, R P is the utilization rate of phosphorus fertilizer (%), P1 is the nitrogen content absorbed by crops in the area where phosphorus fertilizer is applied: F p The amount of phosphorus fertilizer applied.

[0120] In one embodiment:

[0121] Calculate the amount of potassium required for target yield (kg / hm 2 ):

[0122] Soil nutrient potassium supply: K s =K t1 ×K×2.25;

[0123] The amount of potassium required in the current area:

[0124] Among them, U K is the amount of potassium required for the target yield, u is the target yield, which is determined based on the average crop yield in the past three years; K r The amount of potassium required for every 100 kg of crops can be found by searching according to the crop type; K s Potassium supply for soil nutrients, K t1 is the average potassium content in the soil, K is the soil correction factor, and W is the average potassium content in the soil. K is the potassium application rate, C K is the effective content of potassium fertilizer in fertilizer, R K It is the utilization rate of potassium fertilizer.

[0125] In one embodiment, K is the soil effective nutrient correction coefficient, which needs to be prepared in advance by experiments, based on the potassium content absorbed by crops in areas where potassium fertilizer is not applied: The potassium content in the soil of the unfertilized area is K t0 Get, K = K0 ÷ K t0 ×2.25, to obtain the potassium content K0 absorbed by crops in the potassium fertilizer application area and the soil potassium content detection value K in the unfertilized area in multiple different areas of the farmland t0 , with K as the y-axis, K t0 Draw a scatter plot for the x-axis and perform regression analysis to obtain the regression equation y=ax+b, where the a value is the soil calibration coefficient K.

[0126] In one embodiment, R K is the utilization rate of potassium fertilizer (%), K1 is the potassium content absorbed by crops in the area where potassium fertilizer is applied F k The amount of potassium fertilizer applied.

[0127] S7: Generate a variable fertilization prescription map, including: dividing the soil nutrient distribution map into grids of the same size according to the operating range and travel speed of the fertilizer spreader, combining the average soil nutrient content of each grid with the target yield fertilization model, obtaining the required fertilizer amount for a single grid, that is, the fertilization prescription value, and generating a complete fertilization prescription map.

[0128] In one embodiment, the soil nutrient distribution map is divided into several rectangular grids of equal size of 2m*10m, and the average nitrogen, phosphorus and potassium content of a single grid is obtained, substituted into the target yield fertilization model formula, and the fertilizer amount of the corresponding grid is obtained, thereby completing the generation of a variable fertilization prescription map.

[0129] like Figure 4 As shown in the figure, the soil nutrient distribution map is divided into several 2m*10m rectangular grids of equal size, and the average nitrogen, phosphorus and potassium content of a single grid is obtained, and the fertilization formula of the target yield fertilization model is substituted to obtain the fertilization amount of the corresponding grid, so as to complete the generation of the variable fertilization prescription map. The lower left corner of the figure is the color corresponding to the fertilization level. The fertilization level is divided into five levels. Level 1 represents the lack of soil fertility and requires a lot of fertilization; Level 2 represents the slight lack of soil fertility and requires slightly more fertilization; Level 3 represents normal soil fertility and normal fertilization; Level 4 represents the slightly fertile soil fertility and requires a slight reduction in the amount of fertilization; Level 5 represents the fertile soil fertility and requires a small amount of fertilization. The figure shows the distribution of fertilization levels in the field. During the movement of the fertilizer spreader, the location of the field can be determined according to the GPS, the corresponding fertilization level can be obtained, and the variable fertilization of the prescription map can be realized.

[0130] It should be pointed out that although the various steps in the above embodiments are described in a specific order, those skilled in the art can understand that in order to achieve the effects of the present disclosure, different steps do not have to be performed in such an order. They can be performed simultaneously (in parallel) or in other orders. These changes are within the scope of protection of the present disclosure.

[0131] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present disclosure is not limited by the order of the actions described, because according to the present disclosure, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present disclosure.

[0132] The method for improving the speed of generating fertilization prescription maps based on soil spectrum detection described in the present invention is to design a characteristic band between a spectrum and different soil nutrients, and input the reflectance of the characteristic band into the SVR model optimized by the RIME algorithm to quickly obtain soil nutrient data, generate a soil nutrient distribution map, and generate a fertilization prescription map in combination with a corresponding fertilization model, provide fertilization decisions, improve fertilizer utilization, increase the speed of obtaining soil nutrient distribution maps, improve the efficiency of prescription map generation, reduce fertilizer waste, and make it easier to promote the implementation of precision agriculture.

[0133] The method for improving the speed of generating fertilization prescription maps based on soil spectral detection described in the present invention increases the ability of SVR to find optimal parameters through the RIME optimization algorithm, helps SVR to better train data, realizes the estimation of soil nitrogen, phosphorus and potassium nutrients, improves the accuracy of traditional machine learning models, and can realize more accurate detection of soil nutrients.

[0134] To solve the above problems, the present invention also provides a system for quickly generating a fertilization prescription map based on soil spectrum detection, such as Figure 5 As shown, it includes: a characteristic band screening module, a regression model module, a distribution map generation module, a target yield fertilization model module and a prescription map generation module.

[0135] The characteristic band screening module is configured to screen out characteristic bands by inputting the second spectral data and the first soil data into a competitive adaptive reweighted sampling algorithm for feature selection;

[0136] The regression model module is configured to establish a regression model by taking the characteristic band as input and the soil nitrogen, phosphorus and potassium nutrient content as output, and screening out the best parameters of convolution-support vector regression through the frost and ice optimization algorithm;

[0137] The distribution map generation module is configured to obtain soil nutrients according to the regression model and generate a soil nutrient distribution map;

[0138] The target yield fertilization model module is configured to establish a target yield fertilization model; the established target yield fertilization model is stored in the internal memory.

[0139] The prescription map generation module is configured as follows: the soil nutrient distribution map is divided into grids, the fertilizer amount required for a single grid is obtained through the target yield fertilization model, and a complete fertilization prescription map is generated;

[0140] The target yield fertilization model includes: the target yield required nitrogen application model, the target yield required phosphorus application model and the target yield required potassium application model;

[0141] The model of nitrogen application amount required for target yield is as follows:

[0142]

[0143] N s =N t1 ×K×2.25

[0144]

[0145] Among them, U N is the amount of nitrogen required for the target yield, u is the target yield, N r The amount of nitrogen required per 100 kg of crops, N s The nitrogen supply of soil nutrients, N t1 is the average soil nitrogen content detection value, K is the soil correction coefficient, W N is the nitrogen application rate, C N is the effective content of nitrogen fertilizer in fertilizer, R N is the nitrogen fertilizer utilization efficiency.

[0146] Preferably, the method further includes data acquisition, which includes: acquiring soil to be tested, air-drying, grinding, and sieving the soil to be tested, and acquiring first spectral data and first soil data of the processed soil to be tested, wherein the first soil data includes nitrogen, phosphorus, and potassium nutrient content data of the processed soil to be tested.

[0147] Preferably, it also includes a data preprocessing module, and the data preprocessing includes: converting the acquired processed first spectral data of the soil to be tested into an xlsx file, and performing averaging processing and second-order difference preprocessing operations after the conversion is completed to obtain second spectral data.

[0148] Preferably, the soil is sieved with a 0.25 mm standard sieve, the first spectral data is acquired by an ASD ground feature spectrometer, three spectral curves are acquired for a single sample, the sampling interval is 1 nm, and whiteboard correction is performed before each sampling.

[0149] Preferably, converting the obtained first spectral data of the processed soil to be tested into an xlsx file includes: converting the spectral ASD file data into an xlsx file through ENVI software, and the data is stored in the form of: one line is the reflectance of one spectral band, and every three columns are one sample.

[0150] Preferably, the second spectral data and the first soil data are input into a competitive adaptive reweighted sampling algorithm for feature selection, and the characteristic bands are screened out, including: running the competitive adaptive reweighted sampling characteristic band screening algorithm through MATLAB's libpls package, using Monte Carlo cross-validation to determine the number of principal components of the competitive adaptive reweighted sampling, and finally obtaining the best characteristic band.

[0151] As a preferred method, the model of phosphorus application required for target yield is as follows:

[0152]

[0153] P s =P t1 ×K×2.25

[0154]

[0155] Among them, U P is the phosphorus amount required for the target yield, u is the target yield, P r is the amount of phosphorus required per 100 kg of crops, P s The phosphorus supply of soil nutrients, P t1 is the average soil phosphorus content detection value, K is the soil correction coefficient, W P is the phosphorus application rate, C P is the effective content of phosphorus fertilizer in fertilizer; R P is the utilization rate of phosphorus fertilizer;

[0156] The model of potassium application amount required for target yield is as follows:

[0157]

[0158] K s =K t1 ×K×2.25

[0159]

[0160] Among them, U K is the amount of potassium required for the target yield, u is the target yield, K r The amount of potassium required per 100 kg of crops, K s Potassium supply for soil nutrients, K t1 is the average potassium content in the soil, K is the soil correction factor, and W is the average potassium content in the soil. K is the potassium application rate, C K is the effective content of potassium fertilizer in fertilizer, R K It is the utilization rate of potassium fertilizer.

[0161] Preferably, obtaining soil nutrients according to the regression model and generating a soil nutrient distribution map includes: uniformly obtaining a number of soil samples from multiple soil sampling points in the farmland area by a grid sampling method and recording their longitude and latitude coordinates, obtaining the third spectral data of the soil in the corresponding grid by an ASD ground feature spectrometer, obtaining its characteristic band, inputting it into the regression model, quickly obtaining the nitrogen, phosphorus and potassium nutrient contents of the soil samples at each grid sample point, putting the longitude and latitude coordinates of each sample point and the nitrogen, phosphorus and potassium nutrient contents of the soil samples at the sample point into Excel, uploading it to Arcmap software, and generating a soil nutrient distribution map by Kriging spatial interpolation.

[0162] Preferably, the soil nutrient distribution map is divided into grids, and the required fertilizer amount for a single grid is obtained through the target yield fertilization model. Generating a complete fertilization prescription map includes: dividing the soil nutrient distribution map into a number of 2m*10m equal-sized rectangular grids, and obtaining the average nitrogen, phosphorus, and potassium content of a single grid, traversing all grids, and substituting the average nitrogen, phosphorus, and potassium contents of all grids into the target yield fertilization model formula to obtain the required fertilizer amount for all grids, thereby completing the generation of a variable fertilization prescription map.

[0163] In order to solve the above technical problems, the present invention also provides an electronic device, including a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of a method for quickly generating a fertilization prescription map based on soil spectral detection as described above are implemented.

[0164] It is understood that the computer device includes a memory, a processor, and a network interface that are interconnected through a system bus. It should be noted that this embodiment only shows a computer device with component memory, a processor, a network interface, and an operating system, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, it can be understood by those skilled in the art that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a programmable gate array (Field-Programmable Gate Array, FPGA), a digital processor (Digital Signal Processor, DSP), an embedded device, etc.

[0165] Computer devices can be computing devices such as desktop computers, notebooks, PDAs, and cloud servers. Computer devices can interact with users through keyboards, mice, remote controls, touch pads, or voice control devices.

[0166] The memory may be one or more, and may include at least one type of readable storage medium, the readable storage medium including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory may also be an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device. Of course, the memory may also include both an internal storage unit of a computer device and an external storage device thereof. In this embodiment, the memory is generally used to store an operating system and various application software installed on the computer device, such as a program code of a method for quickly generating a fertilization prescription map based on soil spectrum detection, etc. In addition, the memory may also be used to temporarily store various types of data that have been output or are to be output.

[0167] The processor may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor is generally used to control the overall operation of a computer device. In the present embodiment, the processor is used to run the program code stored in the memory or process data, such as running the program code of the method for quickly generating a fertilizer prescription map based on soil spectrum detection.

[0168] The network interface may include a wireless network interface and / or a wired network interface, which is generally used to establish a communication connection between a computer device and other electronic devices.

[0169] The present invention also provides another embodiment, that is, the present invention also provides a readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method for quickly generating a fertilization prescription map based on soil spectral detection as described above are implemented.

[0170] It can be understood that when the computer program is executed by the processor to implement the steps of the method for quickly generating a fertilization prescription map based on soil spectral detection as described above, it can be the entire step flow of the execution method or a partial step flow of the execution method.

[0171] It is understood that the program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0172] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0173] It is not difficult for those skilled in the art to understand that the present invention includes any combination of the invention content and specific implementation methods of the above specification and the various parts shown in the drawings. Due to the limited space and to make the specification concise, the various schemes composed of these combinations are not described one by one. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0174] Although the embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. Those skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention without departing from the principles and purpose of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for quickly generating a fertilization prescription map based on soil spectrum detection, characterized in that: include: The second spectral data of the soil to be tested and the first soil data are subjected to feature selection through a competitive adaptive reweighted sampling algorithm to screen out characteristic bands; the first soil data includes nitrogen, phosphorus and potassium nutrient content data of the soil to be tested; The characteristic bands were used as input and the soil nitrogen, phosphorus and potassium nutrient content was used as output. The optimal parameters of convolution-support vector regression were screened out through the frost optimization algorithm to establish a regression model. Obtain soil nutrients based on the regression model and generate a soil nutrient distribution map; Establish a target yield fertilization model; Divide the soil nutrient distribution map into grids, obtain the required fertilizer amount for a single grid through the target yield fertilization model, and generate a complete fertilization prescription map; The target yield fertilization model includes a nitrogen application amount model required for the target yield; The model of nitrogen application amount required for target yield is as follows: N s =N t1 ×K×2.25 Among them, U N is the amount of nitrogen required for the target yield, u is the target yield, N r The amount of nitrogen required per 100 kg of crops, N s The nitrogen supply of soil nutrients, N t1 is the average soil nitrogen content detection value, K is the soil correction coefficient, W N is the nitrogen application rate, C N is the effective content of nitrogen fertilizer in fertilizer, R N is the nitrogen fertilizer utilization efficiency.

2. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 1, characterized in that: The invention also includes data acquisition, which includes: acquiring soil to be tested, air-drying, grinding, and sieving the soil to be tested, and acquiring first spectral data and first soil data of the processed soil to be tested.

3. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 2 is characterized in that: The method also includes data preprocessing, which includes: converting the acquired processed first spectral data of the soil to be tested into an xlsx file, and performing averaging processing and second-order difference preprocessing operations after the conversion is completed to obtain second spectral data.

4. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 2, characterized in that: The soil was sieved with a 0.25 mm standard sieve. The first spectral data was acquired by an ASD ground feature spectrometer. Three spectral curves were obtained for a single sample with a sampling interval of 1 nm. Whiteboard correction was performed before each sampling.

5. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 3 is characterized in that: The converting of the obtained first spectral data of the processed soil to be tested into an xlsx file comprises: converting the spectral ASD file data into an xlsx file by using ENVI software, wherein the data is stored in a form of: one row is the reflectance of one spectral band, and every three columns are one sample.

6. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 1, characterized in that: The method of performing feature selection on the second spectral data and the first soil data through a competitive adaptive reweighted sampling algorithm to screen out characteristic bands includes: running the competitive adaptive reweighted sampling characteristic band screening algorithm through the libpls package of MATLAB, using Monte Carlo cross-validation to determine the number of principal components of the competitive adaptive reweighted sampling, and finally obtaining the best characteristic band.

7. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 1, characterized in that: The target yield fertilization model also includes: a target yield required phosphorus application model and a target yield required potassium application model; The model of phosphorus application required for target yield is as follows: P s =P t1 ×K×2.25 Among them, U P is the phosphorus amount required for the target yield, u is the target yield, P r is the amount of phosphorus required per 100 kg of crops, P s The phosphorus supply of soil nutrients, P t1 is the average soil phosphorus content detection value, K is the soil correction coefficient, W P is the phosphorus application rate, C P is the effective content of phosphorus fertilizer in fertilizer; R P is the utilization rate of phosphorus fertilizer; The model of potassium application amount required for target yield is as follows: K s =K t1 ×K×2.25 Among them, U K is the amount of potassium required for the target yield, u is the target yield, K r The amount of potassium required per 100 kg of crops, K s Potassium supply for soil nutrients, K t1 is the average potassium content in the soil, K is the soil correction factor, and W is the average potassium content in the soil. K is the potassium application rate, C K is the effective content of potassium fertilizer in fertilizer, R K It is the utilization rate of potassium fertilizer.

8. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 1, characterized in that: The method of obtaining soil nutrients and generating a soil nutrient distribution map according to the regression model includes: uniformly obtaining a number of soil samples from a plurality of soil sampling points in a farmland area by a grid sampling method and recording their longitude and latitude coordinates, obtaining the third spectral data of the soil in the corresponding grid by an ASD ground feature spectrometer, obtaining its characteristic band, inputting it into the regression model, quickly obtaining the nitrogen, phosphorus and potassium nutrient contents of the soil samples at each grid sample point, putting the longitude and latitude coordinates of each sample point and the nitrogen, phosphorus and potassium nutrient contents of the soil samples at the sample point into Excel, uploading it to Arcmap software, and generating a soil nutrient distribution map by Kriging spatial interpolation.

9. The method for quickly generating a fertilization prescription map based on soil spectrum detection according to claim 1, characterized in that: The method of dividing the soil nutrient distribution map into grids, obtaining the required fertilizer amount for a single grid through a target yield fertilization model, and generating a complete fertilization prescription map includes: dividing the soil nutrient distribution map into a number of 2m*10m rectangular grids of equal size, and obtaining the average nitrogen, phosphorus, and potassium content of a single grid, traversing all grids, and substituting the average nitrogen, phosphorus, and potassium content of all grids into the target yield fertilization model formula, respectively, to obtain the required fertilizer amount for all grids, thereby completing the generation of a variable fertilization prescription map.

10. A system for quickly generating a fertilization prescription map based on soil spectrum detection, applicable to a method for quickly generating a fertilization prescription map based on soil spectrum detection according to any one of claims 1 to 9, characterized in that: include: Feature band screening module, regression model module, distribution map generation module, target yield fertilization model module and prescription map generation module, The characteristic band screening module is configured to screen out characteristic bands by inputting the second spectral data and the first soil data into a competitive adaptive reweighted sampling algorithm for feature selection; The regression model module is configured to establish a regression model by taking the characteristic band as input and the soil nitrogen, phosphorus and potassium nutrient content as output, and screening out the best parameters of convolution-support vector regression through the frost and ice optimization algorithm; The distribution map generation module is configured to obtain soil nutrients according to the regression model and generate a soil nutrient distribution map; The target yield fertilization model module is configured to establish a target yield fertilization model; The prescription map generation module is configured to divide the soil nutrient distribution map into grids, obtain the fertilizer amount required for a single grid through the target yield fertilization model, and generate a complete fertilization prescription map; The target yield fertilization model includes: a nitrogen application amount model required for the target yield, a phosphorus application amount model required for the target yield, and a potassium application amount model required for the target yield; The model of nitrogen application amount required for target yield is as follows: N s =N t1 ×K×2.25 Among them, U N is the amount of nitrogen required for the target yield, u is the target yield, N r The amount of nitrogen required per 100 kg of crops, N s The nitrogen supply of soil nutrients, N t1 is the average soil nitrogen content detection value, K is the soil correction coefficient, W N is the nitrogen application rate, C N is the effective content of nitrogen fertilizer in fertilizer, R N is the nitrogen fertilizer utilization efficiency.

Images