Crop growth period determination method, device, electronic device and storage medium

Abstract

The present invention provides a method, device, electronic device and storage medium for determining the growth period of crops, and relates to the field of agricultural remote sensing technology. The method for determining the growth period of crops comprises: obtaining crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height; determining the canopy air temperature difference based on the canopy temperature and the air temperature; judging the growth period of the current crop based on the crop labels, the canopy spectral reflectance, the canopy air temperature difference and the canopy height. The method, device, electronic device and storage medium for determining the growth period of crops provided by the present invention determine the growth period of crops based on near-ground remote sensing data of canopy spectral reflectance, canopy temperature, air temperature and canopy height, making full use of the advantages of existing remote sensing technology in collecting data quickly and efficiently, and quickly and accurately determining the growth period of crops.

Description

Technical Field

[0001] The present invention relates to the field of agricultural remote sensing technology, and in particular to a method, device, electronic equipment and storage medium for determining the growth period of crops. Background Art

[0002] The growth period refers to the growth stage of a crop from sowing to maturity when a typical developmental characteristic appears. It is an important indicator of the growth and development process of the crop, and has guiding significance for the proportional distribution of plant assimilation products to different organs, the formation of yield and quality, and the appropriate time for farmland management. Based on the observation of the growth period, the growth status and development speed of the crop can be grasped. Therefore, the determination of the growth period plays a key role in the fields of crop quality assessment, nutritional diagnosis, yield estimation, low temperature stress, water and fertilizer rationing, variety selection, and farming time monitoring.

[0003] In traditional agriculture, growth period identification is mainly carried out through manual observation. Manual observation relies on observers to enter the field and identify crops according to the growth form and the definition in the relevant production specifications. Due to its reliance on human operation, manual observation has the problems of strong subjectivity, time-consuming and labor-intensive, and large random errors. Summary of the invention

[0004] The invention provides a method, device, electronic device and storage medium for determining the growth period of crops, which are used to solve the problems in the prior art that the determination of the growth period of crops is inefficient and laborious, and has large random errors.

[0005] The present invention provides a method for determining the growth period of crops, comprising:

[0006] Obtain crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height;

[0007] determining a canopy air temperature difference based on the canopy temperature and the air temperature;

[0008] The growth period of the current crop is determined based on the crop label, the canopy spectral reflectance, the canopy air temperature difference and the canopy height.

[0009] According to a crop growth period determination method provided by the present invention, the determining of the current crop growth period based on the crop label, the canopy spectral reflectance, the canopy air temperature difference and the canopy height comprises:

[0010] Determine the canopy leaf area based on the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the canopy leaf area index estimation model corresponding to the crop label;

[0011] Determine the canopy relative chlorophyll content based on the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the canopy relative chlorophyll content estimation model corresponding to the crop label;

[0012] Determining the growth period of the current crop based on the canopy leaf area and the canopy relative chlorophyll content;

[0013] Among them, the canopy leaf area index estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy leaf area sample data; the canopy relative chlorophyll content estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy relative chlorophyll content sample data.

[0014] According to a crop growth period determination method provided by the present invention, the canopy leaf area index estimation model is determined by the least squares method; or,

[0015] The canopy leaf area index estimation model is trained by taking canopy spectral reflectance sample data, canopy air temperature difference sample data and canopy height sample data as input and taking canopy leaf area sample data as output.

[0016] According to a crop growth period determination method provided by the present invention, when the canopy leaf area index estimation model is determined by the least squares method, the canopy leaf area index estimation model of corn is:

[0017] LAI=2.895-24.819x1+196.892x1 2 -83.555x2+1056.576x2 2 +1.949x3-0.307x3 2 -0.012x4 -0.022x4 2

[0018] In the formula, the corn variety is Nongda 86, the corn planting density is 60 cm row spacing, 20 cm plant spacing, and the seeding rate is 37.5 kg / ha. LAI represents the canopy leaf area index, and x1 represents the red light reflectance R in the 650 mm band. NIR , x2 represents the near-infrared reflectivity R in the 810mm band RED , x3 represents the canopy temperature difference △T, and x4 represents the canopy height.

[0019] According to a crop growth period determination method provided by the present invention, the canopy relative chlorophyll content estimation model is determined by using a regression analysis method; or,

[0020] The canopy relative chlorophyll content estimation model is trained by taking canopy spectral reflectance sample data, canopy air temperature difference sample data and canopy height sample data as input and taking canopy relative chlorophyll content sample data as output.

[0021] According to a crop growth period determination method provided by the present invention, the determining of the current crop growth period based on the canopy leaf area and the canopy relative chlorophyll content comprises:

[0022] The crop accumulated temperature is determined based on the canopy leaf area, the canopy relative chlorophyll content and a fuzzy algorithm, and the growth period of the current crop is determined based on the crop accumulated temperature.

[0023] A method for determining the growth period of a crop provided by the present invention further comprises:

[0024] Obtain canopy leaf area and canopy relative chlorophyll content,

[0025] The determining of the current crop growth period based on the crop label, the canopy spectral reflectance, the canopy air temperature difference and the canopy height specifically includes:

[0026] Determine the current crop growth period based on the crop label, the canopy leaf area, the canopy relative chlorophyll content, the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the crop accumulated temperature calculation model;

[0027] Among them, the crop accumulated temperature calculation model is constructed based on canopy leaf area sample data, canopy relative chlorophyll content sample data, canopy spectral reflectance sample data, canopy temperature difference sample data, and canopy height sample data.

[0028] The present invention also provides a crop growth period determination device, comprising:

[0029] an acquisition unit, the acquisition unit being used to acquire crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height,

[0030] A determination unit is used to determine the canopy air temperature difference based on the crop label, the canopy temperature and the air temperature, and to judge the growth period of the current crop based on the canopy spectral reflectance, the canopy air temperature difference and the canopy height.

[0031] 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 any of the above-mentioned methods for determining the growth period of crops are implemented.

[0032] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods for determining the growth period of crops are implemented.

[0033] The crop growth period determination method, device, electronic device and storage medium provided by the present invention determine the crop growth period based on near-ground remote sensing data of canopy spectral reflectance, canopy temperature, air temperature and canopy height, making full use of the advantages of existing remote sensing technology in collecting data conveniently and efficiently to quickly and accurately determine the crop growth period. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 This is one of the flow charts of the crop growth period determination method provided by the present invention;

[0036] Figure 2 This is the second flow chart of the method for determining the crop growth period provided by the present invention;

[0037] Figure 3 This is a diagram showing the effect of the canopy leaf area index estimation model determined based on the least squares method provided by the present invention;

[0038] Figure 4 This is a diagram showing the effect of the canopy leaf area index estimation model determined based on the BP neural network provided by the present invention;

[0039] Figure 5 It is a diagram showing the effect of the canopy relative chlorophyll content estimation model determined based on different machine learning methods provided by the present invention;

[0040] Figure 6 It is a framework diagram of the fuzzy learning method provided by the present invention;

[0041] Figure 7 The crop growth period surface distribution map obtained based on fuzzy rules provided by the present invention;

[0042] Figure 8 It is a structural schematic diagram of a crop growth period determination device provided by the present invention;

[0043] Fig. 9 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] Combine the following Figure 1-Figure 7 The crop growth period determination method of the present invention is described.

[0046] Figure 1 This is one of the flow charts of the crop growth period determination method provided by the embodiment of the present invention. Figure 1 As shown in the figure, the methods for determining the crop growth period include:

[0047] Step 100, obtaining crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height;

[0048] Step 200, determining the canopy air temperature difference based on the canopy temperature and the air temperature;

[0049] Step 300, determining the growth period of the current crop based on the crop label, canopy spectral reflectance, canopy air temperature difference and canopy height.

[0050] Among them, crop labels include crop varieties and planting densities. Among them, crop varieties are specific varieties under crop types. Taking corn as an example, crop varieties are specific corn varieties, such as Agricultural University 86. The planting density is based on the density setting at the time of sowing. When establishing the canopy leaf area index estimation model and the canopy relative chlorophyll content estimation model, they need to be constructed separately for different crop varieties and different planting densities, so as to form multiple models corresponding to different planting densities of the same crop variety and multiple models corresponding to the same planting density of different crop varieties in the model library. When calling, the corresponding model is selected according to the crop variety and planting density. Crop labels are input manually.

[0051] The canopy spectral reflectance includes the red light reflectance R NIR and near infrared reflectivity R RED, obtained through a dual-channel multispectral SRS-NDVI sensor. In addition, other types of spectral detection equipment can be used to detect the spectral reflectance of the canopy. The relative chlorophyll content of the canopy is measured by the SPAD-502 chlorophyll meter. The canopy height is measured by ultrasonic ranging sensors such as the TSPC-15 ultrasonic ranging sensor or sensors such as radar. The canopy temperature is measured by the SI-111 infrared temperature sensor, and the air temperature is obtained by measuring the field automatic weather station or by using the air temperature detection unit on the infrared sensor. The canopy air temperature difference is obtained by subtracting the canopy temperature from the air temperature. It should be noted that the air temperature does not need to be detected in real time. When it is necessary to determine the crop growth period, it is sufficient to obtain the current canopy temperature and air temperature of the crop.

[0052] In one embodiment of the present invention, the crop canopy spectral reflectance, canopy temperature, air temperature, canopy height, canopy leaf area and canopy relative chlorophyll content are obtained, the canopy air temperature difference is determined based on the canopy temperature and the air temperature, the crop accumulated temperature is determined based on the canopy spectral reflectance, canopy air temperature difference, canopy height, canopy leaf area and canopy relative chlorophyll content, and then the crop growth period is determined according to the corresponding relationship between the crop accumulated temperature of the current crop and the crop growth period. Among them, the canopy leaf area is measured manually or by a leaf area meter, and the canopy relative chlorophyll content is obtained by a SPAD sensor or a chemical analysis method. The crop accumulated temperature is calculated by a crop accumulated temperature calculation model determined by canopy spectral reflectance sample data, canopy temperature sample data, air temperature sample data, canopy height sample data, canopy leaf area sample data and canopy relative chlorophyll content sample data. It can be understood that different crop varieties and different planting densities correspond to different crop accumulated temperature calculation models. When the crop variety and planting density are determined, the crop growth period determination method does not need to monitor the air temperature of the crop in real time. It only needs to obtain the current canopy temperature, air temperature, canopy height and canopy spectral reflectance of the crop when making judgments to determine the crop growth period.

[0053] In another embodiment of the present invention, the canopy leaf area and the canopy relative chlorophyll content are estimated based on the canopy spectral reflectance, the canopy air temperature difference and the canopy height, respectively, the crop accumulated temperature is determined based on the estimated canopy leaf area and the canopy relative chlorophyll content, and the crop growth period is determined based on the accumulated temperature. The canopy leaf area estimation model is constructed based on the canopy spectral reflectance sample data, the canopy air temperature difference sample data, the canopy height sample data and the canopy leaf area sample data through least squares method or BP neural network algorithm, extreme learning machine and other machine learning methods. The canopy relative chlorophyll content estimation model is constructed based on the canopy spectral reflectance sample data, the canopy air temperature difference sample data, the canopy height sample data and the canopy relative chlorophyll content sample data through BP neural network algorithm or extreme learning algorithm or support vector set and other machine learning methods or determined by regression analysis. The model library needs to be constructed for different crop varieties and different planting densities. After the crops are planted, the planting density and crop variety are determined, and the corresponding model can be retrieved for subsequent operations.

[0054] The crop growth period determination method provided in the embodiment of the present invention determines the crop growth period based on the near-ground remote sensing data of canopy spectral reflectance, canopy temperature, air temperature and canopy height, making full use of the advantages of existing remote sensing technology in collecting data conveniently and efficiently, and quickly and accurately determining the crop growth period.

[0055] Figure 2 This is the second flow chart of the crop growth period determination method provided by the embodiment of the present invention. Figure 2 As shown, based on the above embodiment, judging the current crop growth period based on canopy spectral reflectance, canopy air temperature difference and canopy height includes:

[0056] Step 210, determining the canopy leaf area based on the canopy spectral reflectance, canopy air temperature difference, canopy height and a canopy leaf area index estimation model corresponding to the crop label;

[0057] The canopy relative chlorophyll content is determined based on the canopy spectral reflectance, canopy air temperature difference, canopy height and the canopy relative chlorophyll content estimation model corresponding to the crop label;

[0058] Step 220, determining the current crop growth period based on the canopy leaf area and the canopy relative chlorophyll content.

[0059] Among them, the canopy leaf area index estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy leaf area sample data; the canopy relative chlorophyll content estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy relative chlorophyll content sample data.

[0060] In an optional embodiment, the canopy leaf area index estimation model is determined using the least squares method.

[0061] The following specifically introduces the construction of the canopy leaf area index estimation model using corn as the experimental crop.

[0062] The test site is located at the Tongzhou Experimental Station of China Agricultural University (39°41'59″N, 116°41'01″E). The test corn was managed with normal water and fertilizer management, and the variety was Nongda 86. The summer corn was sown on June 23, 2019, with a planting density of 60 cm in row and 20 cm in plant, and a sowing rate of 37.5 kg / ha. The sampling growth stages were selected from the jointing stage (V6), small trumpet stage (V9), large trumpet stage V12 (V12), tasseling stage (VT), and early filling stage (R), as shown in Table 1.

[0063] It should be noted that the model construction process is similar for different crop varieties and different planting densities. The following model construction process is introduced in detail using the crop variety Nongda 86, the planting density of 60 cm row spacing and 20 cm plant spacing as an example.

[0064] Table 1 Sampling time of each growth period

[0065]

[0066] The following test data were all measured in a clear, cloudless and windless environment. The specific sampling days are shown in Table 1, and the sampling period was between 10:00 and 14:00.

[0067] The spectral reflectance of corn canopy includes red light reflectance R NIR and near infrared reflectivity R RED The spectral reflectance of the corn canopy obtained in the experiment is the red light reflectance R in the 650nm band. NIR and near-infrared reflectance in the 810nm band. When the least squares method is used to construct the canopy leaf area index estimation model, its evaluation indicators include the determination coefficient R 2 , mean square error MSE, root mean square error RMSE and normalized root mean square error nRMSE. The calculation formulas for each evaluation index are as follows:

[0068]

[0069] Where: SSR is the regression sum of squares, SSE is the residual sum of squares, P i is an estimated value, Q i is the measured value, is the average value of the measured value, and n is the number of samples.

[0070] The canopy height, canopy temperature difference △T, and red light reflectance R NIR and near infrared reflectivity R REDThe four types of experimental parameters were fitted with the canopy leaf area index by multiple factors. The canopy air temperature difference sample data was determined based on the collected canopy temperature sample data and air temperature sample data. The collected canopy height sample data, the calculated canopy air temperature difference △T sample data, the collected red light reflectance sample data and the near-infrared reflectance sample data were used as training sets and validation sets. The multivariate nonlinear canopy leaf area index estimation model was established using the least squares method. The canopy leaf area index estimation model for corn was determined as follows:

[0071] LAI=2.895-24.819x1+196.892x1 2 -83.555x2+1056.576x2 2 +1.949x3-0.307x3 2 -0.012x4 -0.022x4 2

[0072] In the formula, LAI represents the canopy leaf area index, and x1 represents the red light reflectance R in the 650nm band. NIR , x2 represents the near-infrared reflectivity R in the 810nm band RED , x3 represents the canopy temperature difference (△T), and x4 represents the canopy height. In the following formula, x1, x2, x3, and x4 have the same meaning and will not be repeated here.

[0073] The red light reflectance R NIR and near infrared reflectivity R RED The determined two-factor canopy leaf area index estimation model is:

[0074] LAI=3.2+69.418x1-321.319x1 2 -289.926x2+4930.458x2 2

[0075] The red light reflectance R NIR , near infrared reflectivity R RED The three-factor canopy leaf area index estimation model determined by the canopy temperature difference △T is:

[0076] LAI=5.112+17.633x1-64.698x1 2 -206.427x2+3246.880x2 2 -0.08x3 -0.03x3 2

[0077] The red light reflectance R NIR , near infrared reflectivity R RED The three-factor canopy leaf area index estimation model determined by the canopy height is:

[0078] LAI=1.441-12.498x1+133.540x1 2 -84.06x2+1019.68x2 2 +2.725x4-0.479x4 2

[0079] The four-factor, two-factor and two three-factor canopy leaf area index estimation models were compared, as shown in Table 2.

[0080] Table 2 Multivariate nonlinear LAI estimation model

[0081]

[0082] The comparison between the estimated values ​​calculated by the four-factor canopy leaf area index estimation model and the measured values ​​of the canopy leaf area index in the validation set is shown in the figure below: Figure 3 As shown in Figure 2, the fitting effect is good. Specifically, the determination coefficient R of the four-factor canopy leaf area index estimation model is 2 =0.917, RMSE=0.38, nRMSE=8.9%. Analysis of Table 2 shows that the canopy leaf area index estimation model determined by four factors has better estimation ability.

[0083] In another optional embodiment, the canopy leaf area index estimation model is trained with canopy spectral reflectance sample data, canopy air temperature difference sample data and canopy height sample data as input, and canopy leaf area sample data as output.

[0084] Specifically, the sample library data is divided into a training set and a validation set in a ratio of 4:1 or 3:1. The canopy leaf area index estimation model is constructed based on the training set, and the canopy height calibration model effect is verified based on the validation set. For example, the canopy leaf area index estimation model is determined using machine learning methods such as deep learning or BP neural network.

[0085] Still taking the sample data used in the least squares modeling as the sample library, another method for constructing the canopy leaf area index estimation model is introduced by taking the BP neural network model as an example. First, based on the red light reflectance R NIR and near infrared reflectivity R RED A variety of vegetation indices are calculated. For example, the ratio vegetation index RVI, the difference vegetation index DVI, the normalized vegetation index NDVI, the improved ratio vegetation index MSR, the renormalized vegetation index RDVI, the soil adjusted vegetation index SAVI, the optimized soil adjusted vegetation index OSAVI, the modified soil adjusted vegetation index MASVI, the enhanced vegetation index EVI2 and the enhanced modified soil adjusted vegetation index MASVI2. The calculation formulas of each vegetation index are shown in Table 3.

[0086] Table 3 Vegetation index involved in this experiment

[0087]

[0088] Substitute the above vegetation index as input parameters into the model construction process, and the number of elements in the two-factor, three-factor, and four-factor models are 10, 11, and 12 respectively. According to the requirements for BP neural network construction, the number of neurons in the input layer and output layer should be equal to the number of input parameters and output parameters respectively. At the same time, it can be determined by empirical formulas that the number of nodes in the hidden layer is 4 to 14. Substitute all data into MATLAB software for training. Select the case with the largest R and the smallest SME in the results to determine the number of nodes, and finally select the two-factor and three-factor models. 1 , three factors 2 The number of hidden layer nodes of the four-factor BP neural network is 14, 12, 14 and 14 respectively. In order to ensure the stability of modeling and avoid overfitting, the embodiment of the present invention also introduces a four-fold cross-validation accuracy test method, so that the four types of BP neural network models can achieve a relatively stable prediction level.

[0089] The above BP neural network model was verified using the validation set. As shown in Table 4, the estimated results were compared with the measured values, including canopy height, canopy air temperature difference, red light reflectance R NIR and near infrared reflectivity R RED The BP neural network model with four types of input parameters has better LAI estimation ability. The comparison between the estimated values ​​calculated by the four-factor canopy leaf area index estimation model determined by the BP neural network and the measured values ​​of the canopy leaf area index in the validation set is shown in the figure. Figure 4 shown.

[0090] Table 4 Evaluation results of each model

[0091]

[0092] In a specific embodiment, the canopy relative chlorophyll content estimation model is trained with canopy spectral reflectance sample data, canopy air temperature difference sample data and canopy height sample data as input, and canopy relative chlorophyll content sample data as output. For example, it is trained using BP neural network algorithm or extreme learning algorithm or support vector set and other methods. Among them, the canopy relative chlorophyll content estimation model can also be determined by regression analysis, and its analysis process is similar to the above-mentioned process of determining the canopy leaf area index estimation model by the least squares method, and will not be repeated. The following takes machine learning as an example to introduce the construction process of the canopy relative chlorophyll content estimation model.

[0093] The sample library used for model construction is the same as above. The sample library data is divided into a training set and a validation set according to a 4:1 or 3:1 allocation ratio. The canopy relative chlorophyll content estimation model is built based on the training set, and the accuracy of the canopy relative chlorophyll content estimation model is tested based on the validation set.

[0094] Since the expression of canopy relative chlorophyll content SPAD will affect the estimation accuracy of the model, the canopy relative chlorophyll content, i.e., SPAD value, was subjected to three mathematical transformations, namely lg(SPAD), 1 / SPAD, and 1 / lg(SPAD). Among them, lg is the logarithmic value of the canopy relative chlorophyll content with base 10; 1 / SPAD is the reciprocal of the canopy relative chlorophyll content; 1 / lg is the reciprocal of the logarithmic value of the canopy relative chlorophyll content with base 10. The models were established using three machine learning algorithms: BP neural network, support vector machine (SVM), and extreme learning machine (ELM). In addition to canopy height and canopy air temperature difference (△T), the input parameters required for modeling are also combined with R NIR , R RED Six vegetation indices, namely, ratio vegetation index (RVI), difference vegetation index (DVI), normalized difference vegetation index (NDVI), modified ratio vegetation index (MSR), renormalized difference vegetation index (RDVI), and enhanced vegetation index 2 (EVI2), were selected as input parameters based on the correlation characteristics with chlorophyll content. See Table 3 for details.

[0095] The evaluation indicators obtained during the modeling process include: determination coefficient R 2 , mean error MBE, mean square error MSE, standard root mean square error nRMSE, fitting index d. Among them, MBE is the average error between the measured value and the simulated value, which can be used to judge whether the model simulation value overestimates or underestimates the measured value. When MBE>0, it means that the model simulation overestimates the measured value; when MBE<0, it means that the model simulation underestimates the measured value. Fit index d (0≤d≤1) is a descriptive statistic, a dimensionless statistical evaluation index, which can be used for statistical comparison between output variables of different units. When d=1, it means that the model simulation value is very consistent with the measured value, while d=0 indicates that the simulation result is very poor; when d≥0.9, it means that the fit between the simulated value and the measured value is very good; when 0.8≤d<0.9, it means that the model simulation performance is good; when 0.7≤d<0.8, it means that the model simulation performance is medium; when d<0.7, it means that the model simulation performance is poor. The specific calculation formula is as follows.

[0096]

[0097]

[0098] Where: M i is the actual measured value; Si is the model prediction value; The actual measured value is the mean, and n is the number of samples.

[0099] The number of input parameters can determine that the number of input nodes of the BP neural network is 8. According to the empirical formula, the number of hidden layer nodes of this network is calculated to be in the range of 3 to 13. According to the selection rule of maximum correlation coefficient r and minimum SME, the number of hidden layer nodes of SPAD, 1 / SPAD, lg(SPAD) and 1 / lg(SPAD) inversion models are 8, 8, 9 and 9 respectively. In order to ensure the stability of modeling and avoid overfitting, the accuracy test method of four-fold cross validation is introduced here, so that the difference between the estimated value of the canopy relative chlorophyll content estimation model determined by the BP neural network algorithm and the actual value is small, and the estimation performance is better.

[0100] Based on the validation set, the canopy relative chlorophyll content estimation models determined by BP neural network, support vector machine (SVM) and extreme learning machine (ELM) were verified. The validation results are as follows: Figure 5 The evaluation results of the canopy relative chlorophyll content estimation model determined by BP neural network, support vector machine (SVM) and extreme learning machine (ELM) are shown in Table 5. Figure 5 As shown in Table 5, the estimation model with 1 / lg(SPAD) as the output parameter has better prediction effect than lg(SPAD), 1 / SPAD, and SPAD. Based on extreme learning machine (ELM), BP neural network, and support vector machine (SVM), better canopy relative chlorophyll content estimation models can be determined. Preferably, the canopy relative chlorophyll content estimation model is determined by extreme learning machine.

[0101] Table 5 Evaluation of various machine learning models

[0102]

[0103]

[0104] Based on the above embodiment, determining the current crop growth period based on the canopy leaf area and the canopy relative chlorophyll content specifically includes:

[0105] The crop accumulated temperature is determined based on the canopy leaf area, canopy relative chlorophyll content and fuzzy algorithm, and the current crop growth period is determined based on the crop accumulated temperature.

[0106] Since the heat accumulation required by crops to complete a certain growth stage must reach a certain accumulated temperature, the accumulated temperature calculation method is used to implement the crop growth period determination based on fuzzy algorithm in MATLAB software. Figure 6As shown in the figure, the canopy leaf area index LAI and the canopy relative chlorophyll content SPAD are used as input parameters, and the crop growth period CGP is used as the output parameter. The fuzzy controller uses the Mamdami fuzzy logic system with high stability to build the fuzzy rule base; the "and", "or", "implication", "comprehensive", and "clarification" in the fuzzy logic respectively select the widely used "smaller", "larger", "smaller", "and", and "area center method". Combining the sample library with the field test data, the input and output variables and the corresponding actual domains can be obtained. The actual domains corresponding to the input and output variables are shown in Table 6, where the unit of accumulated temperature (≥10℃) at each growth stage in CGP is (℃·d). According to the sensitivity of the canopy leaf area index LAI, the canopy relative chlorophyll content SPAD and the crop growth period CGP data, the Gaussian, triangle and triangle membership functions are selected respectively.

[0107] Table 6 Variable domain

[0108]

[0109] The embodiment of the present invention formulates 25 fuzzy rules based on the field corn growth and development law and the test database, which are implemented as a whole according to the law that "with the increase of crop accumulated temperature, the canopy leaf area index LAI and the canopy relative chlorophyll content SPAD gradually increase". The specific fuzzy logic rules are shown in Table 7. Based on this rule, the output CGP surface distribution is as follows Figure 7 shown.

[0110] Table 7 Fuzzy logic rules table

[0111]

[0112] It should also be noted that the LAI and SPAD value ranges that only appear when abnormal conditions occur in the field observation area are not considered in the fuzzy logic rule table. The specific abnormal situation rules are shown in Table 8.

[0113] Table 8 Abnormal situation rules

[0114]

[0115]

[0116] After determining the fuzzy controller structure, fuzzy logic system, membership function and logic rules, the LAI and SPAD values ​​can be substituted into the fuzzy controller as input parameters, and then the accumulated temperature value can be output. Finally, the crop growth period can be determined based on the accumulated temperature value.

[0117] It should be noted that if an ultrasonic ranging sensor is used to collect the canopy height, the collected canopy height is corrected, and the canopy height calibration value obtained after the correction is used to determine the crop growth period. That is, the embodiment of the present invention also provides a method for determining the crop growth period, which includes:

[0118] Obtain crop tags, canopy spectral reflectance, canopy temperature, air temperature, and canopy height measured by ultrasonic ranging sensors;

[0119] Determine the canopy air temperature difference based on canopy temperature and air temperature;

[0120] determining a canopy height calibration value based on the canopy height and a corresponding canopy height calibration model;

[0121] The current crop growth period is determined based on crop labels, canopy spectral reflectance, canopy air temperature difference and canopy height calibration values.

[0122] Wherein, determining the canopy height calibration value based on the canopy height and the canopy height calibration model specifically includes:

[0123] Get the normalized difference index of vegetation;

[0124] Determine the canopy height calibration value based on the vegetation normalized index, canopy height and canopy height calibration model corresponding to the crop label;

[0125] Among them, the canopy height calibration model is constructed based on the vegetation normalized index sample data, canopy height measurement value sample data and the actual value of canopy height.

[0126] Crop labels are manually input, and the vegetation normalization index is obtained through a spectral camera or NDVI sensor, which can reflect the planting density of crops. Canopy height measurements are collected through ultrasonic ranging sensors. Different crop labels have different canopy height calibration models. According to the crop label, the canopy height calibration model corresponding to the crop label is called, and the canopy height measurement value and the vegetation normalization index are substituted into the canopy height calibration model to calculate the corrected canopy height calibration value.

[0127] The canopy height calibration model of corn was obtained by multivariate nonlinear regression analysis:

[0128]

[0129] Where CH is the actual value of the canopy height, x1 is the measured value of the canopy height, and x2 is the normalized vegetation index.

[0130] Figure 8 Schematic diagram of the structure of the crop growth period determination device provided by the present invention. Figure 8 As shown, the crop growth period determination device comprises:

[0131] The acquisition unit 810 is used to acquire crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height.

[0132] The determination unit 820 is used to determine the canopy air temperature difference based on the crop label, the canopy temperature and the air temperature and to judge the growth period of the current crop based on the canopy spectral reflectance, the canopy air temperature difference and the canopy height.

[0133] Fig. 9 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Fig. 9 As shown, the electronic device may include: a processor 910, a communication interface 920, a memory 930 and a communication bus 940, wherein the processor 910, the communication interface 920 and the memory 930 communicate with each other through the communication bus 940. The processor 910 may call the logic commands in the memory 930 to execute the following method: obtaining crop tags, canopy spectral reflectance, canopy temperature, air temperature and canopy height; determining the canopy air temperature difference based on the canopy temperature and the air temperature; judging the growth period of the current crop based on the crop tags, canopy spectral reflectance, canopy air temperature difference, canopy temperature and canopy height.

[0134] In addition, the logic commands in the above-mentioned memory 930 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several commands to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk.

[0135] An embodiment of the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it is implemented to execute the methods provided in the above embodiments, for example, including: obtaining crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height; determining the canopy air temperature difference based on the canopy temperature and the air temperature; judging the growth period of the current crop based on crop labels, canopy spectral reflectance, canopy air temperature difference, canopy temperature and canopy height.

[0136] The present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the crop canopy height measurement calibration method provided by the above methods, which includes: obtaining crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height; determining the canopy air temperature difference based on the canopy temperature and the air temperature; judging the current crop's growth period based on the crop label, canopy spectral reflectance, canopy air temperature difference, canopy temperature and canopy height.

[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining the growth period of crops, characterized in that: include: Obtain crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height; determining a canopy air temperature difference based on the canopy temperature and the air temperature; Determine the canopy leaf area based on the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the canopy leaf area index estimation model corresponding to the crop label; Determine the canopy relative chlorophyll content based on the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the canopy relative chlorophyll content estimation model corresponding to the crop label; Determining the growth period of the current crop based on the canopy leaf area and the canopy relative chlorophyll content; The canopy leaf area index estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy leaf area sample data; the canopy relative chlorophyll content estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy relative chlorophyll content sample data; The canopy leaf area index estimation model is determined by the least squares method; the canopy leaf area index estimation model for corn is: <h2 style=";text-align:left;direction:ltr">LAI=2.895-24.819x1+196.892x1<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> -83.555x2+1056.576x2<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +1.949x3- 0.307x3 2 -0.012x4-0.022x4 2 In the formula, the corn variety is Nongda 86, the corn planting density is 60 cm in row and 20 cm in plant, the seeding rate is 37.5 kg / ha, LAI represents the canopy leaf area index, and x1 represents the red light reflectance R in the 650 nm band. NIR , x2 represents the near-infrared reflectivity R in the 810nm band RED , x3 represents the canopy temperature difference △T, and x4 represents the canopy height.

2. The method for determining the crop growth period according to claim 1, characterized in that: The canopy relative chlorophyll content estimation model is determined by regression analysis; or, The canopy relative chlorophyll content estimation model is trained by taking canopy spectral reflectance sample data, canopy air temperature difference sample data and canopy height sample data as input and taking canopy relative chlorophyll content sample data as output.

3. The method for determining the crop growth period according to any one of claims 1 to 2, characterized in that: Determining the growth period of the current crop based on the canopy leaf area and the canopy relative chlorophyll content comprises: The crop accumulated temperature is determined based on the canopy leaf area, the canopy relative chlorophyll content and a fuzzy algorithm, and the growth period of the current crop is determined based on the crop accumulated temperature.

4. The method for determining the crop growth period according to claim 1, characterized in that: Also includes: Obtain canopy leaf area and canopy relative chlorophyll content, The determining of the current crop growth period based on the crop label, the canopy spectral reflectance, the canopy air temperature difference and the canopy height specifically includes: Determine the current crop growth period based on the crop label, the canopy leaf area, the canopy relative chlorophyll content, the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the crop accumulated temperature calculation model; Among them, the crop accumulated temperature calculation model is constructed based on canopy leaf area sample data, canopy relative chlorophyll content sample data, canopy spectral reflectance sample data, canopy temperature difference sample data, and canopy height sample data.

5. A device for determining the growth period of crops, characterized in that: include: an acquisition unit, the acquisition unit being used to acquire crop labels, canopy spectral reflectance, canopy temperature, air temperature and canopy height, A determination unit, the determination unit is used to determine the canopy leaf area based on the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the canopy leaf area index estimation model corresponding to the crop label; determine the canopy relative chlorophyll content based on the canopy spectral reflectance, the canopy air temperature difference, the canopy height and the canopy relative chlorophyll content estimation model corresponding to the crop label; and determine the current crop growth period based on the canopy leaf area and the canopy relative chlorophyll content; The canopy leaf area index estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy leaf area sample data; the canopy relative chlorophyll content estimation model is determined based on canopy spectral reflectance sample data, canopy air temperature difference sample data, canopy height sample data and canopy relative chlorophyll content sample data; The canopy leaf area index estimation model is determined by the least squares method; the canopy leaf area index estimation model of corn is: <h2 style=";text-align:left;direction:ltr">LAI=2.895-24.819x1+196.892x1<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> -83.555x2+1056.576x2<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +1.949x3- 0.307x3 2 -0.012x4-0.022x4 2 In the formula, the corn variety is Nongda 86, the corn planting density is 60 cm in row and 20 cm in plant, the seeding rate is 37.5 kg / ha, LAI represents the canopy leaf area index, and x1 represents the red light reflectance R in the 650 nm band. NIR , x2 represents the near-infrared reflectivity R in the 810nm band RED , x3 represents the canopy temperature difference △T, and x4 represents the canopy height.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the crop growth period determination method according to any one of claims 1 to 4 are implemented.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the crop growth period determination method as claimed in any one of claims 1 to 4 are implemented.

Images