CN110321627A - Merge the photosynthetic rate prediction technique of leaf photosynthesis potential - Google Patents
Merge the photosynthetic rate prediction technique of leaf photosynthesis potential Download PDFInfo
- Publication number
- CN110321627A CN110321627A CN201910576601.XA CN201910576601A CN110321627A CN 110321627 A CN110321627 A CN 110321627A CN 201910576601 A CN201910576601 A CN 201910576601A CN 110321627 A CN110321627 A CN 110321627A
- Authority
- CN
- China
- Prior art keywords
- photosynthetic
- photosynthetic rate
- model
- data
- potential
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000000243 photosynthetic effect Effects 0.000 title claims abstract description 128
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000029553 photosynthesis Effects 0.000 title claims abstract description 12
- 238000010672 photosynthesis Methods 0.000 title claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000012706 support-vector machine Methods 0.000 claims abstract description 22
- 238000012360 testing method Methods 0.000 claims abstract description 18
- 238000005286 illumination Methods 0.000 claims abstract description 14
- 230000008569 process Effects 0.000 claims abstract description 12
- 230000002068 genetic effect Effects 0.000 claims description 18
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 claims description 15
- 229910002092 carbon dioxide Inorganic materials 0.000 claims description 14
- 239000001569 carbon dioxide Substances 0.000 claims description 8
- 108090000623 proteins and genes Proteins 0.000 claims description 7
- 239000013598 vector Substances 0.000 claims description 7
- 210000000349 chromosome Anatomy 0.000 claims description 6
- 230000004907 flux Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 230000035790 physiological processes and functions Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 4
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000002474 experimental method Methods 0.000 abstract description 12
- 230000004927 fusion Effects 0.000 abstract description 3
- 238000013461 design Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 abstract 1
- 238000005192 partition Methods 0.000 abstract 1
- 241000196324 Embryophyta Species 0.000 description 30
- 244000061458 Solanum melongena Species 0.000 description 8
- 235000002597 Solanum melongena Nutrition 0.000 description 6
- 230000007613 environmental effect Effects 0.000 description 6
- 238000012795 verification Methods 0.000 description 6
- 238000005457 optimization Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000012010 growth Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 241000894007 species Species 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 229930002875 chlorophyll Natural products 0.000 description 1
- 235000019804 chlorophyll Nutrition 0.000 description 1
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000004300 dark adaptation Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000001506 fluorescence spectroscopy Methods 0.000 description 1
- 239000005556 hormone Substances 0.000 description 1
- 229940088597 hormone Drugs 0.000 description 1
- 230000004298 light response Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 239000000575 pesticide Substances 0.000 description 1
- 238000006552 photochemical reaction Methods 0.000 description 1
- 230000008635 plant growth Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000027756 respiratory electron transport chain Effects 0.000 description 1
- 238000005316 response function Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000011282 treatment Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N2021/635—Photosynthetic material analysis, e.g. chrorophyll
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Chemical & Material Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Genetics & Genomics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Physiology (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Computational Linguistics (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Apparatus Associated With Microorganisms And Enzymes (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The photosynthetic rate prediction technique of leaf photosynthesis potential, the identical plant photosynthesis test of design other parameters in addition to light levels are merged, the plant to different lighting processes generates difference, randomly selects the discrepant plant of physiological status as experiment sample；Measure temperature, the CO of difference blade2Net Photosynthetic Rate under concentration and intensity of illumination Nested conditions, and the dark fluorescence parameter of blade is recorded, as sample data；To normalized is done on sample data different dimensions, each dimension data is made to be in an order of magnitude, and partition testing collection and training set together；The photosynthetic rate prediction model of fusion leaf photosynthesis potential is established using Support Vector Machines for Regression algorithm；Using the prediction model, the photosynthetic rate of fusion leaf photosynthesis potential is predicted, the present invention provides most important theories basis for the control accurate of facilities environment and technology is realized.
Description
Technical Field
The invention belongs to the technical field of intelligent agriculture, relates to photosynthetic rate prediction, and particularly relates to a photosynthetic rate prediction method fusing photosynthetic potential of leaves.
Background
Photosynthesis is the key to plant growth, provides an energy source for plant carbon accumulation, and the yield of crops is closely related to the photosynthetic accumulation of the crops. The photosynthesis rate of plants is influenced by the environment and the state of the plants, and the external environment including temperature, atmospheric carbon dioxide concentration, photosynthetic light quantum flux density and the like has important influence on the photosynthesis of crops. Meanwhile, the photosynthetic potentials of the plant leaves at different periods and different leaf positions are greatly different, so that the reason for different photosynthetic rates is that the inherent photosynthetic potentials of the plant leaves are different. The photosynthetic potential of the crops is fused, the external environmental factors of the crops are coupled, a photosynthetic rate model based on the coupling of the photosynthetic potential of the crops and the facility environment multifactor is established, and the photosynthetic rate of the crops can be reflected more accurately.
Numerous scholars have conducted extensive studies on the plant photosynthetic rate model, and k.w.brown proposed atmospheric CO2The illumination intensity-photosynthetic rate model under the concentration predicts the CO fixed by the leaves2Without ever considering CO2Concentration, etc. on the rate of photosynthesis. Jingz et al establish photosynthetic rate response functions of the effects of various environmental factors on photosynthetic rate, reflect the effects of various factors on photosynthetic rate to a certain extent, but do not consider the coupling relationship of multi-environmental factors on plant photosynthetic rate. Zhang Haihui etc. takes account of the difference of photosynthetic capacity of plant leaves at different leaf positions, establishes a photosynthetic rate model fusing the leaf positions, further improves the accuracy of the photosynthetic rate model, but does not consider the difference of photosynthetic capacity of leaves at the same leaf position in different states.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a photosynthetic rate prediction method, a photosynthetic experiment under nesting of facility environmental factors of different fluorescent leaves is designed, the photosynthetic potential of the leaves of a plant is fused, a photosynthetic rate prediction model is established based on a regression type support vector machine optimized by an improved genetic algorithm after data are preprocessed, accurate photosynthetic rate prediction can be realized by utilizing the model, a uniform model can be provided for the photosynthetic rates of the leaves in different growth states, and a foundation is established for regulation and control of the facility agricultural environmental factors.
In order to achieve the purpose, the invention adopts the technical scheme that:
a photosynthetic rate prediction method fusing photosynthetic potential of leaves comprises the following steps:
s1, designing plant photosynthesis tests with the same parameters except the illumination level, wherein the physiological states of leaves of plants to be treated by different illuminations are different, and randomly selecting plants with obvious differences in the physiological states of the leaves as test samples;
step S2, measuring the temperature and CO of the difference blade2Net photosynthetic rate under nested conditions of concentration and illumination intensity, and recording dark fluorescence parameters of leaves as sample data;
step S3, normalization processing is carried out on different dimensions of the sample data, so that the data of all dimensions are in one order of magnitude, and a test set and a training set are divided;
step S4, establishing a photosynthetic rate prediction model fusing the photosynthetic potential of the leaves by utilizing a regression type support vector machine algorithm;
and step S5, predicting the photosynthetic rate of the photosynthetic potential of the fused leaves by using the prediction model.
In the step S1, the illumination level is 1-6, and other parameters are day/night light cycle, day/night temperature, day/night relative humidity and CO2And (4) concentration.
In step S3, the normalization interval is [0,1], and the formula is:
wherein x is the original data, xminIs the minimum value of all raw data，xmaxIs the maximum of all the raw data.
In step S3, 80% of the sample data is randomly divided into a training set, and the remaining 20% of the sample data is divided into a test set, where the input characteristic of the ith sample is xi＝(xi (1),xi (2),xi (3),xi (4),xi (5)) The output label is yi，xi (1),xi (2),xi (3),xi (4),xi (5)The light quantum flux density, the carbon dioxide concentration, the temperature, the minimum fluorescence parameter Fo and the maximum fluorescence parameter Fm, y of the ith sampleiThe photosynthetic rate of the ith sample.
In step S4, the improved genetic algorithm is used to solve the regularization parameter c and the kernel function parameter g that are optimal for the model, the index range of c is set to [0.01,50], the number of genes is 6, the index range of g is set to [0.01,5], the number of genes is 5, the coding mode is set to binary coding, and the coding formula is:
the decoding formula is:
wherein b is the coded binary string, m is the number of characters of the binary string taken by the chromosome, a is the decimal number to be coded, aminFor coding the space minimum decimal number, amaxIs the maximum decimal number of the coding space;
using decoded c and g as parameters, training a regression type support vector machine by using training set data, calculating a decision coefficient of test set data, using the decision coefficient as chromosome fitness, carrying out population elimination and selection by using a roulette wheel, simulating manual intervention in a species selection process in formed offspring, replacing an offspring lowest fitness individual by a parent maximum fitness individual, forming new offspring by crossing, variation and selection operations of the offspring, and carrying out iterative evolution to finally converge to optimal solution parameters c and g. The coefficient of variation is selected to be 0.2, the cross coefficient is selected to be 0.8, and the optimal solution parameters c and g can be obtained by iterative evolution for 50 generations.
In step S4, the modeling process is as follows:
s4.1, reading in modeling data and carrying out normalization pretreatment;
s4.2, selecting 80% of sample data as a training set, and selecting modeling parameters including a regularization parameter c, a kernel function and a kernel function parameter g;
s4.3, solving regression hyperplane parameters according to the KKT condition to obtain a photosynthetic rate model;
step S4.4, substituting the test set data into the model to obtain fitting data, calculating a fitting error, and if the fitting error does not meet the requirement, repeating the steps S4.2 to S4.4 until the fitting error meets the requirement, so as to obtain a final model f (x) ═ wTAnd x + v, wherein w is an input coefficient, and v is the intercept of the model in each axis.
In model training, the model introduces a relaxation variable epsiloniForming the functional soft interval, the regression support vector machine problem can be transformed into:
the target is as follows:
constraint of f (x)i)-yi≤ε+εi,εi≥0,i＝1,2,…,M
Wherein, yiThe photosynthetic rate of the ith sample is epsilon, an insensitive loss function is determined by the support vector; epsiloniAndis a relaxation variable; and M is the total number of model training samples, and for the non-linear problem,and mapping the input vector to a higher-dimensional space through a kernel function, converting the nonlinear problem into a linear problem to solve, constructing a regression hyperplane, and finally mapping the regression hyperplane back to the original space to convert the regression hyperplane into a nonlinear hypersurface.
And (3) selecting a Gauss kernel function for dimension transformation, wherein the Gauss kernel function is in the form of:
wherein X is a sample to be transformed, XpThe center of the p-th kernel function is defined, g is a kernel function parameter, and finally, the model is learned by combining kernel skills as follows:
wherein,αiis the lagrange multiplier in the solution process.
Compared with the prior art, the invention has the beneficial effects that:
1) a photosynthetic rate prediction model based on plant photosynthetic potential is established, plant dark fluorescence parameters Fo and Fm are used as input factors when the model is established, and the phenomenon that the traditional photosynthetic rate prediction model cannot accurately predict the photosynthetic rates of leaves in different states is effectively avoided.
2) The regression support vector machine parameters with the highest sample data fitting accuracy can be found in a short time by using the genetic algorithm, and the genetic algorithm after the genetic mode is improved can effectively avoid the situation that the optimal individual is lost in the evolution process, so that the optimization convergence speed is higher.
3) A regression type support vector machine is used for establishing a photosynthetic rate prediction model, the model establishment is determined by a small amount of support vectors, the modeling time is accelerated, and the generalization capability and the fitting precision are effectively improved by the soft interval and kernel function design. After the model is established, 5-fold cross validation is used, and the generalization capability of the validation model to unknown data is strong.
Drawings
FIG. 1 is a flow chart of the photosynthetic rate prediction model building with the leaf photosynthetic potential of the present invention fused.
FIG. 2 is a flow chart of the improved genetic algorithm of the present invention.
FIG. 3 is a flow chart of the regression support vector machine algorithm of the present invention.
FIG. 4 is a photosynthetic curve of leaves with different fluorescence values.
FIG. 5 is a photosynthetic rate error surface based on different model parameters according to the present invention.
FIG. 6 is a graph of the evolution of the genetic algorithm of the present invention.
FIG. 7 is a diagram illustrating the correlation between measured values and fitting values of the light combination rate in the model verification according to the present invention.
FIG. 8 is a distribution of the calculated data and the raw data for the fluorescence-free parametric model.
Detailed Description
The embodiments of the present invention will be described in detail below with reference to the drawings and examples.
The invention relates to a photosynthetic rate prediction method for fusing leaf photosynthetic potential, which evaluates the photosynthetic potential of plants by using chlorophyll dark fluorescence parameters as indexes for measuring the photosynthetic capacity of leaves. The different dark fluorescence parameters have different meanings, the maximum fluorescence Fm parameter can reflect the electron transfer condition of the PSII reaction center, the initial fluorescence Fo parameter can represent the light energy radiation of the PSII reaction center which does not participate in the photochemical reaction, and the initial fluorescence Fo parameter and the PSII reaction center can reflect the activity of the PSII reaction center and represent the photosynthetic potential of the plant leaf.
The invention adopts a regression support vector machine algorithm to establish a photosynthetic rate model fusing the photosynthetic potential of plants, and obtains the optimal support vector machine parameters by an improved genetic algorithm. And (4) taking 80% of the experimental samples as training sets to train the models, taking the remaining 20% of samples as test sets to test the generalization ability of the models, and performing model evaluation.
As shown in fig. 1, taking eggplant as an example, the process of the prediction method of the present invention is as follows:
1 materials and methods
1.1 test materials and methods
The experiment is carried out in an agricultural Internet of things key laboratory (34 degrees and 07 ' 39 degrees in northern latitude, 107 degrees and 59 ' 50 degrees in east longitude, 107 degrees and 59 ' 50 degrees in elevation, 648 meters) of northwest rural agricultural science and technology university in Shaanxi province from 10 months to 2019 months in 2018 (internal experiment, confidentiality state). The material for experiment was "Long eggplant 305", the substrate used for planting was Pindstrupestrate (Pindstrupestrate), and the experimental eggplant plants were cultivated in RGL-P500D-CO produced by Hefedbatchet corporation2Carbon dioxide artificial box. The incubator is set with photoperiod of day/night: 14/10 hours, set temperature day/night: 25 ℃/16 ℃, relative humidity set day/night: 60%/50%, setting CO2The concentration was 400. mu. mol/mol. The 54 eggplant seedlings are randomly and evenly divided into 6 groups, the groups are respectively cultivated in an incubator with the illumination grade of 1-6, the rest environmental factors and the incubator are managed uniformly, watering is unified, and no hormone or pesticide is sprayed. And (3) when the eggplant plants subjected to different light treatments have differences, randomly selecting the plants with obvious differences in each incubator as experimental samples.
The temperature and CO of the different blades are designed through experiments2And measuring and recording the dark fluorescence parameters of the leaves at the net photosynthetic rate under the nested condition of concentration and illumination intensity. In the experiment, the fluorescence data of the leaves of eggplant was measured using a MINI-PAM-II fluorometer manufactured by WALZ, Germany, and the net photosynthetic rate of the leaves of eggplant was measured using an LI-6800 photosynthesizer manufactured by LI-COR, USA. The net photosynthetic rate measurement experiment sets the light quantum flux density gradient of the photosynthetic apparatus to 1500120010008006003001506030150 mu mol/m2S, setting leaf chamber CO2The concentration gradient was 13001000700400. mu. mol/mol, and the temperature gradient in the leaf chamber was set at 3531272319 ℃. The data measurement experiment starts from 27 days 2 and 3 days 21 in 2019, and four leaves with larger differences are randomly selected to carry out the experiment on the same day. Dark adaptation was first performed for 20 minutes for each leaf using dark-adapted leaf clamps, followed by measurement of dark fluorescence parameters and net photosynthetic rate to obtain 1295 sets of data for modeling.
1.2 model building method
The process for establishing the regression type support vector machine accurate photosynthetic rate model based on the improved genetic algorithm optimization parameters mainly comprises sample data processing, model parameter selection and model construction.
1.2.1 data preprocessing
Normalization processing is needed to be carried out on different dimensions of sample data before a support vector machine is trained, so that the data of all dimensions are in the same order of magnitude, sample imbalance caused by overlarge data difference is avoided, and finally a model deviates from an accurate hyperplane. The normalized interval is [0,1], and the formula is:
wherein x is the original data, xminIs the minimum value, x, of all the raw datamaxIs the maximum of all the raw data.
After the data set is normalized, the data set is randomly divided into 80% of training set and 20% of testing set for model training. Wherein the input characteristic of the ith sample is xi＝(xi (1),xi (2),xi (3),xi (4),xi (5)) The output label is yi，xi (1),xi (2),xi (3),xi (4),xi (5)The light quantum flux density, the carbon dioxide concentration, the temperature, the minimum fluorescence parameter Fo and the maximum fluorescence parameter Fm, y of the ith sampleiThe photosynthetic rate of the ith sample.
1.2.2 model parameter selection
The genetic algorithm is an algorithm which quickly converges to an optimal value point by simulating the natural selection and evolution of species in nature, and is widely applied to the actual problem solving by effectively preventing the local optimization and the efficient nonlinear solving. The photosynthetic rate model parameters c and g of the photosynthetic capacity of the fused eggplant leaves based on the regression type support vector machine are solved by an improved genetic algorithm.
The algorithm sets the index range of c as [0.01,50] and the number of genes as 6, sets the index range of g as [0.01,5] and the number of genes as 5.
Setting the encoding mode as binary encoding, wherein the encoding formula is as follows:
the decoding formula is:
wherein b is the coded binary string, m is the number of characters of the binary string taken by the chromosome, a is the decimal number to be coded, aminFor coding the space minimum decimal number, amaxIs the largest decimal number in the coding space.
And (3) training a regression type support vector machine by using the decoded c and g as parameters and training set data, calculating a decision coefficient of a prediction set data set, using the decision coefficient as chromosome fitness, and performing population elimination and selection by using a roulette plate. In the formed offspring, the manual intervention of the species selection process is simulated, and the highest fitness individual of the parent is substituted for the lowest fitness individual of the offspring. And forming new filial generations through crossing, mutation and selection operations, and carrying out iterative evolution to finally converge to the optimal solution parameters c and g. Because the gene diversity in the population is reduced by manual intervention operation, a suitably large population variation coefficient should be selected, the selection variation coefficient is 0.2, the cross coefficient is 0.8, and the optimal solution parameters c and g are obtained by iterative evolution for 50 generations. The flow chart of the improved genetic algorithm is shown in fig. 2.
1.2.3 photosynthetic rate prediction model establishment for leaf photosynthetic potential fusion
For a given training sample { (x)1,y1),(x2,y2),…,(xm,ym) The regression-type support vector machine can train a regression model f (x) wTx + v, where w is the input coefficient and v is the model intercept at each axis, let f (x)i) And yiAs close as possible, the modeling flow is shown in fig. 3.
In model training, the model introduces a relaxation variable epsiloniForming a soft interval of the function, and supporting the regression model toThe problems of measuring machines are converted into:
the target is as follows:
constraint of f (x)i)-yi≤ε+εi,εi≥0,i＝1,2,…,m
Wherein c is a regularization parameter, which influences the model accuracy; epsilon is an insensitive loss function and is determined by a support vector; epsiloniAndis the relaxation variable.
For nonlinear problems, a support vector machine converts the nonlinear problems into linear problems to solve through a kernel technique, and the specific method is to map input vectors into a higher-dimensional space through a kernel function, convert the nonlinear problems into linear problems to solve, construct a regression hyperplane, and finally map the regression hyperplane back to an original space to convert the regression hyperplane into a nonlinear hypersurface. The Gauss kernel function has the characteristics of parameter change and unchanged model complexity, and the Gauss kernel function is selected for the model to carry out dimension transformation. The Gauss kernel function is of the form:
wherein X is a sample to be transformed, XpIs the center of the p-th kernel function, and g is a kernel function parameter, which influences the kernel function form.
Combining the nuclear skills, the model is finally learned to be:
wherein,αiis a solving processLagrange multiplier of (1).
2 results and discussion
2.1 comparison of the photoresponse curves of the leaves with different fluorescence parameters
The photosynthetic rate of the leaves with different fluorescence parameters is different under the same external environment. FIG. 4 shows the light response curves of leaves with different fluorescence parameters under the condition of consistent temperature and carbon dioxide concentration, and the difference is obvious and can be attributed to the difference of the photosynthetic potential of the leaves. As shown in the figure, the smaller Fo value and the larger Fm value of the photosynthetic potential of the leaf are larger, and the photosynthetic capacity is larger under the same external conditions. And the external environments required by the leaves with different fluorescence values at the same photosynthetic rate are different, which has important significance for guiding the accurate regulation and control of facility environment.
2.2 model Performance and validation results analysis
The accuracy of the photosynthetic rate model trained by the regression support vector machine algorithm on data prediction is closely related to the parameters c and g, and different prediction errors are obtained by different parameters. The mean square error curve of the model for the unknown data prediction is shown in fig. 5, which shows that there are relatively optimal parameters, so that the prediction error of the model for the unknown data can be minimized.
The optimal parameter value is obtained by optimizing with an improved genetic algorithm, and the obtained genetic evolution curve is shown in fig. 6. The decision coefficients of the model prediction dataset and the actual dataset are increased as the genetic algebra increases, and finally converge to 0.993, and the root mean square error when the optimal parameters (c 41.07, g 3.30) are obtained is 0.246 μmol · m-2s-1The data can be fitted with higher accuracy.
In order to verify the generalization capability of the model, a 5-fold cross verification method is adopted, sample data is randomly divided into five parts, the model is set by using parameters c and g obtained by an optimization algorithm, one part is sequentially selected as a verification set, and the remaining four parts are used for training the model, so that five models under the parameters c and g obtained by optimization are finally obtained. The performance of each model is verified through verification set data to obtain five groups of verification data. The maximum coefficient of the decision of the model prediction data and the minimum coefficient of the decision of the verification data are 0.992 and 0.988, and the approximation degrees of the two are shown in FIG. 7, which shows that the model has excellent generalization capability
The leaves in different physiological states have larger photosynthetic potential difference, when the photosynthetic potential in the leaves is not introduced as a photosynthetic rate influencing factor and only external environmental factors are used as model input, different leaf photosynthetic rate models have larger difference, and a nested experimental data set is difficult to fit an accurate photosynthetic rate model. Therefore, in the process of building and researching the photosynthetic rate model by the scholars in the past, plants with strictly consistent growth vigor must be adopted for the experiment, the same leaf position is required to be used for the position measured by the experimental data, and the built model is poor in universality.
Dark fluorescence parameters representing plant photosynthetic potential are removed from the experimental sample, and only external PPFD and CO are used2The concentration and the atmospheric temperature are input parameters, the photosynthetic rate is used as an output parameter, and a photosynthetic rate model is established under the same condition. The coefficient of the model for determining unknown prediction set data is 0.5909, and the root mean square error is 14.4937 mu mol m-2s-1The distribution of the calculated value and the real value of the photosynthetic rate model is shown in fig. 8, which is difficult to be applied in practice.
Therefore, the photosynthetic rate model established by fusing the light and the potential in the plant by adopting the support vector machine optimized by the improved genetic algorithm has excellent fitting effect and generalization capability, and can realize accurate prediction of the photosynthetic rate of the plant leaves in different states.
In summary, the invention is based on the principle that the photosynthetic rate of the crop is not only related to the external environment, but also related to the state of the plant, and the photosynthetic rate model fused with the photosynthetic potential of the plant can more accurately predict the photosynthetic rate of the plant, the concentration of carbon dioxide, the illumination intensity and the temperature in the greenhouse environment are main external factors influencing the photosynthetic rate, the photosynthetic potential of the plant leaves is an internal factor influencing the photosynthesis of the plant, and the photosynthetic potential of the plant is reflected by the dark fluorescence parameters of the plant. Experiments are designed to obtain the photosynthetic rate of the leaves with different dark fluorescence parameters under the nesting of temperature, carbon dioxide concentration and illumination light quantum flux density, and 1295 groups of data are obtained. Establishing a crop photosynthetic rate model by using a regression type support vector machine algorithm according to the obtained dataMatching the optimal parameters by using an improved genetic algorithm to obtain a model with the data decision coefficient of 0.993 and the root mean square error of 0.246 mu mol.m for an unknown prediction set-2s-1. Based on the obtained model, 5-fold cross validation is carried out on all data, the maximum coefficient of determination is 0.992, and the minimum coefficient of determination is 0.988, and the model is proved to have excellent generalization capability. A photosynthetic rate model fusing internal and external influence factors of crops is established for the crops with different photosynthetic potentials, and an important theoretical basis and technical realization are provided for the accurate regulation and control of facility environment.
The photosynthetic rate prediction model based on the photosynthetic potential of the leaves can provide theoretical basis for accurate facility environment regulation and control of plants in different growth states, obviously, the model can be expanded to different crops, and the facility crop efficiency is improved.
Claims (9)
1. The photosynthetic rate prediction method for the photosynthetic potential of the fused leaves is characterized by comprising the following steps of:
s1, designing plant photosynthesis tests with the same parameters except the illumination level, wherein the physiological states of leaves of plants to be treated by different illuminations are different, and randomly selecting plants with obvious differences in the physiological states of the leaves as test samples;
step S2, measuring the temperature and CO of the difference blade2Net photosynthetic rate under nested conditions of concentration and illumination intensity, and recording dark fluorescence parameters of leaves as sample data;
step S3, normalization processing is carried out on different dimensions of the sample data, so that the data of all dimensions are in one order of magnitude, and a test set and a training set are divided;
step S4, establishing a photosynthetic rate prediction model fusing the photosynthetic potential of the leaves by utilizing a regression type support vector machine algorithm;
and step S5, predicting the photosynthetic rate of the photosynthetic potential of the fused leaves by using the prediction model.
2. The method for predicting photosynthetic rate of photosynthetic potential of fused leaf according to claim 1,in the step S1, the illumination level is 1-6, and other parameters are day/night light cycle, day/night temperature, day/night relative humidity and CO2And (4) concentration.
3. The method for predicting photosynthetic rate of photosynthetic potential of the fused leaf according to claim 1, wherein in step S3, the normalized interval is [0,1], and the formula is:
wherein x is the original data, xminIs the minimum value, x, of all the raw datamaxIs the maximum of all the raw data.
4. The method for predicting photosynthetic rate of photosynthetic potential of fused leaf according to claim 1 or 3, wherein in step S3, 80% of sample data are randomly divided into training set, and the remaining 20% of sample data are divided into testing set, wherein the input characteristic of the ith sample is xi＝(xi (1),xi (2),xi (3),xi (4),xi (5)) The output label is yi，xi (1),xi (2),xi (3),xi (4),xi (5)The light quantum flux density, the carbon dioxide concentration, the temperature, the minimum fluorescence parameter Fo and the maximum fluorescence parameter Fm, y of the ith sampleiThe photosynthetic rate of the ith sample.
5. The photosynthetic rate prediction method of photosynthetic potential of the fused leaf blade of claim 1, wherein in step S4, the optimized regularization parameter c and kernel function parameter g of the model are solved by a modified genetic algorithm, an index range of c is set to [0.01,50], the number of genes is 6, an index range of g is set to [0.01,5], the number of genes is 5, the coding mode is set to binary coding, and the coding formula is as follows:
the decoding formula is:
wherein b is the coded binary string, m is the number of characters of the binary string taken by the chromosome, a is the decimal number to be coded, aminFor coding the space minimum decimal number, amaxIs the maximum decimal number of the coding space;
using decoded c and g as parameters, training a regression type support vector machine by using training set data, calculating a decision coefficient of test set data, using the decision coefficient as chromosome fitness, carrying out population elimination and selection by using a roulette wheel, simulating manual intervention in a species selection process in formed offspring, replacing an offspring lowest fitness individual by a parent maximum fitness individual, forming new offspring by crossing, variation and selection operations of the offspring, and carrying out iterative evolution to finally converge to optimal solution parameters c and g.
6. The method for predicting photosynthetic rate of photosynthetic potential of fused leaf as claimed in claim 5, wherein the optimal solution parameters c and g are obtained through iterative evolution for 50 generations with a coefficient of variation of 0.2 and a cross coefficient of 0.8.
7. The photosynthetic rate prediction method of photosynthetic potential of the fused leaf as claimed in claim 5 or 6, wherein the modeling process of step S4 is as follows:
s4.1, reading in modeling data and carrying out normalization pretreatment;
s4.2, selecting 80% of sample data as a training set, and selecting modeling parameters including a regularization parameter c, a kernel function and a kernel function parameter g;
s4.3, solving regression hyperplane parameters according to the KKT condition to obtain a photosynthetic rate model;
step S4.4, substituting the test set data into the model to obtain fitting data, calculating a fitting error, and if the fitting error does not meet the requirement, repeating the steps S4.2 to S4.4 until the fitting error meets the requirement, so as to obtain a final model f (x) ═ wTAnd x + v, wherein w is an input coefficient, and v is the intercept of the model in each axis.
8. The method for predicting photosynthetic rate of photosynthetic potential of fused leaf according to claim 7, wherein the model introduces a relaxation variable ε during model trainingiForming the functional soft interval, the regression support vector machine problem can be transformed into:
the target is as follows:
constraint of f (x)i)-yi≤ε+εi,εi≥0,i＝1,2,…,M
Wherein, yiThe photosynthetic rate of the ith sample is epsilon, an insensitive loss function is determined by the support vector; epsiloniAndis a relaxation variable; and M is the total number of model training samples, for a nonlinear problem, an input vector is mapped to a higher-dimensional space through a kernel function, the nonlinear problem is converted into a linear problem to be solved, a regression hyperplane is constructed, and finally the regression hyperplane is mapped back to an original space and converted into a nonlinear hypersurface.
9. The photosynthetic rate prediction method of photosynthetic potential of fused leaf as claimed in claim 8, wherein a Gauss kernel function is selected for dimensional transformation, the Gauss kernel function being in the form of:
wherein X is a sample to be transformed, XpThe center of the p-th kernel function is defined, g is a kernel function parameter, and finally, the model is learned by combining kernel skills as follows:
wherein,αiis the lagrange multiplier in the solution process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910576601.XA CN110321627B (en) | 2019-06-28 | 2019-06-28 | Photosynthetic rate prediction method for photosynthetic potential of fused leaves |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910576601.XA CN110321627B (en) | 2019-06-28 | 2019-06-28 | Photosynthetic rate prediction method for photosynthetic potential of fused leaves |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110321627A true CN110321627A (en) | 2019-10-11 |
| CN110321627B CN110321627B (en) | 2022-10-14 |
Family
ID=68120598
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910576601.XA Active CN110321627B (en) | 2019-06-28 | 2019-06-28 | Photosynthetic rate prediction method for photosynthetic potential of fused leaves |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110321627B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112735511A (en) * | 2021-01-13 | 2021-04-30 | 西北农林科技大学 | Cold damage cucumber PSII potential activity prediction method based on QGA-SVR |
| CN114002194A (en) * | 2021-10-22 | 2022-02-01 | 安徽袁粮水稻产业有限公司 | Method for identifying photosynthetic capacity of improved rice plant |
| CN114814099A (en) * | 2022-04-25 | 2022-07-29 | 南京农业大学 | Photosynthesis prediction method based on grape leaf shape |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654203A (en) * | 2015-12-31 | 2016-06-08 | 西北农林科技大学 | Cucumber whole-course photosynthetic rate predicting model based on support vector machine, and establishing method |
| CN105678405A (en) * | 2015-12-31 | 2016-06-15 | 西北农林科技大学 | Modeling method for cucumber GA-RBF photosynthetic rate prediction model integrated with stomatal conductance |
| CN107341734A (en) * | 2017-06-06 | 2017-11-10 | 浙江大学 | A kind of method for building up of the protected crop seedling growth forecast model based on physiological parameter |
| WO2019090878A1 (en) * | 2017-11-09 | 2019-05-16 | 合肥工业大学 | Analog circuit fault diagnosis method based on vector-valued regularized kernel function approximation |
-
2019
- 2019-06-28 CN CN201910576601.XA patent/CN110321627B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654203A (en) * | 2015-12-31 | 2016-06-08 | 西北农林科技大学 | Cucumber whole-course photosynthetic rate predicting model based on support vector machine, and establishing method |
| CN105678405A (en) * | 2015-12-31 | 2016-06-15 | 西北农林科技大学 | Modeling method for cucumber GA-RBF photosynthetic rate prediction model integrated with stomatal conductance |
| CN107341734A (en) * | 2017-06-06 | 2017-11-10 | 浙江大学 | A kind of method for building up of the protected crop seedling growth forecast model based on physiological parameter |
| WO2019090878A1 (en) * | 2017-11-09 | 2019-05-16 | 合肥工业大学 | Analog circuit fault diagnosis method based on vector-valued regularized kernel function approximation |
Non-Patent Citations (1)
| Title |
|---|
| 王辉等: "不同遮阴处理对香草兰光合作用及花芽分化的影响", 《福建农业学报》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112735511A (en) * | 2021-01-13 | 2021-04-30 | 西北农林科技大学 | Cold damage cucumber PSII potential activity prediction method based on QGA-SVR |
| CN112735511B (en) * | 2021-01-13 | 2023-11-03 | 西北农林科技大学 | Cold damage cucumber PSII potential activity prediction method based on QGA-SVR |
| CN114002194A (en) * | 2021-10-22 | 2022-02-01 | 安徽袁粮水稻产业有限公司 | Method for identifying photosynthetic capacity of improved rice plant |
| CN114814099A (en) * | 2022-04-25 | 2022-07-29 | 南京农业大学 | Photosynthesis prediction method based on grape leaf shape |
| CN114814099B (en) * | 2022-04-25 | 2023-09-12 | 南京农业大学 | Photosynthesis prediction method based on grape leaf shape |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110321627B (en) | 2022-10-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110321627B (en) | Photosynthetic rate prediction method for photosynthetic potential of fused leaves | |
| CN107145941B (en) | Method for dynamically acquiring light demand quantity in real time based on optimal light quality and photon flux density | |
| CN112906298B (en) | Blueberry yield prediction method based on machine learning | |
| CN112735511B (en) | Cold damage cucumber PSII potential activity prediction method based on QGA-SVR | |
| CN107329511A (en) | Based on the vegetable aquaculture luminous environment high efficiency regulatory method and system between suitable root warm area | |
| Hu et al. | Model for tomato photosynthetic rate based on neural network with genetic algorithm | |
| CN105654203A (en) | Cucumber whole-course photosynthetic rate predicting model based on support vector machine, and establishing method | |
| CN111915062B (en) | Greenhouse crop water demand regulation and control method with water utilization rate and photosynthesis rate being coordinated | |
| CN105678405B (en) | Modeling method for cucumber GA-RBF photosynthetic rate prediction model integrated with stomatal conductance | |
| CN113361761A (en) | Short-term wind power integration prediction method and system based on error correction | |
| CN109214591B (en) | Method and system for predicting aboveground biomass of woody plant | |
| CN108614601A (en) | A kind of facility luminous environment regulation and control method of fusion random forests algorithm | |
| CN115453868B (en) | Full-growth-period light intensity regulation and control method based on tomato light response difference characteristics | |
| CN105446142A (en) | Greenhouse CO2 gas fertilizer increasing method, device and system | |
| CN107220672A (en) | Acquisition methods between a kind of suitable warm area based on crop demand | |
| CN110414729A (en) | The potential maximum photosynthetic capacity prediction technique of plant based on characteristic wavelength | |
| CN110569605A (en) | Non-glutinous rice leaf nitrogen content inversion model method based on NSGA2-ELM | |
| Wang et al. | Cotton growth model under drip irrigation with film mulching: A case study of Xinjiang, China | |
| CN113962819A (en) | Method for predicting dissolved oxygen in industrial aquaculture based on extreme learning machine | |
| Liu et al. | Prediction model of photosynthetic rate based on SOPSO-LSSVM for regulation of greenhouse light environment | |
| Shang et al. | Research on intelligent pest prediction of based on improved artificial neural network | |
| Pao et al. | How does structure matter? Comparison of canopy photosynthesis using one-and three-dimensional light models: a case study using greenhouse cucumber canopies | |
| Zuxing et al. | A Prediction Model of Forest Preliminary Precision Fertilization Based on Improved GRA‐PSO‐BP Neural Network | |
| CN112364710B (en) | Plant electric signal classification and identification method based on deep learning algorithm | |
| Jamshidi et al. | High throughput mathematical modeling and multi-objective evolutionary algorithms for plant tissue culture media formulation: Case study of pear rootstocks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |