Research On Prediction And Evaluation Of Light Environment Of Cultivated Ginseng Under Forest

Panax Ginseng C.A.Meyer belongs to Araliaceae plant, which has thousands of yearsof history in our country and is known as the king of grass bouquet. For a long time, themain ginseng production mode of our country is that wild ginseng excavation and gardenginseng cultivation. Ginseng once had a wide distribution in the history of our country,however, as the development of wild resources and the excessive excavation, wild ginsengproduction declined year by year and resources had been largely dried up and on the brinkof extinction. Garden ginseng cultivation has a long history with wide planting area andyield, but the traditional method that deforestation to ginseng cultivation and returning landfor ginseng cultivation to forestry, caused serious ecological destruction; At the same time itchanged the soil structure, easy to cause soil harden, and will produce great influence ondiversity of species resources and ecology recovery, and also existing some problems thatpesticides cause a low content of active ingredients in ginseng and excessive pesticideresidues, and poor processing quality, which leads to high yield and low prices, exportrestraints, etc. Therefore, the method of ginseng cultivation under forests which can makefull use of forest land and does not destroy resources, improve the ecological benefit isgradually taken seriously. Though its quality is not equal of wild ginseng, but much higherthan of garden ginseng and favored by the market.As shade plants species, ginseng has a feature that favoring cool and moist climate,dislike bright light and high temperature, and cold-resistant, so in the process of planting, itis very strict with light conditions. When too much light, ginseng growth is restrained,leaves are susceptible to damage, photosynthesis is reduced; But when insufficient light,plants are small and thin, has poor growth. It can be thought of understory lightenvironment is key factors to undergrowth development in the process of ecosystem.Therefore study of light environment and its evaluation and prediction has become a reasonable guide to the ginseng cultivation under forests. The paper took the lightenvironment of cultivation of ginseng under forests as the research object, on the basis ofresearch results at the early stage of the research group, put forward a kind of nonlinearFourier analysis method for solving the problem of adaptive processing of the measuredsignals of light environment. At the same time using the machine learning and patternrecognition theory, constructed the model of net photosynthetic rate for factors analysisbased on partial least-square regression (PLS) algorithm and the model of netphotosynthetic rate prediction based on adaptive neural fuzzy inference (ANFIS), andthrough data sampling of typical test sample areas, the feasibility and effectiveness of themodels is verified. Then the light environmental data acquisition system of cultivation ofginseng under forests was designed based on the adaptive algorithm of data processing, andfinally constructed the prediction and evaluation system of light environment of cultivatedginseng under forest based on the MATLAB platform which can obtain the suitability indexof evaluation of regional cultivation for light environment of ginseng cultivation and asimportant guidance basis of ginseng cultivation.The main research contents of this paper are as follows:(1) Aiming at the random disturbance problem in the process of the related physicalquantities measurement of ginseng cultivation light environment, some typical dataprocessing algorithms are compared, such as fast Fourier transform, add window Fouriertransform, wavelet analysis, etc., and then nonlinear fast Fourier decomposition algorithmwas put forward for non-stationary signal processing, discussed the nonlinear convergenceproperties of Fourier expansion, then using the sparse approximation theory to construct thefast discrete algorithm to be used to calculate sparse nonlinear coefficient, realized ddimensional function nonlinear sparse. For convenience of practical application, this papergives the algorithm steps, and select two different Sobolev regularity of function withdifferent parameters for analysis, a numerical example analysis shows that the nonlinearFourier expansion approximation accuracy is significantly higher than the linear Fourierexpansion. Through the noise reduction processing applied with Fourier expansionalgorithm to the simulation signal containing random noise and net photosynthetic ratemeasured data, the results demonstrated the effectiveness and feasibility of the algorithm.(2) Aiming at the light environment characteristics of ginseng cultivation under forest, using the algorithm advantage of intelligent computing in solving adaptive modelprediction, through the analysis and discussion of pattern recognition theory and methods,and machine learning which applied to solve the problems of regression and classification,analyzing and comparing the least squares (LS), partial least squares (PLS), support vectormachine (SVM), gaussian process (GP), and other typical algorithm characteristics andworking mechanism of machine learning algorithms. The method of PLS of whichrequirements and component extraction and modeling steps were discussed, and using thealgorithm to implement component extraction and analysis, then the net photosynthetic rateof ginseng PLS multi-factor analysis model was obtained. The regression coefficient ofeach variable distribution and its correlation with the dependent variable were also analyzed,of which the results showed that the model has well prediction effect.(3) Through the analysis of physiological indicators of Pinus Koraiensis, usingstepwise regression analysis method, the prediction models were established whichincluded the canopy breadth growth model, the crown length growth model as well as thebasic tree high growth model. Amount of statistical bias model tests showed that the abovetree growth models have high accuracy. Based on PLS method, analyzed the influencingfactors of net photosynthetic rate of ginseng under forest by principal component extractionand weight analysis, then got the multi-factor analysis model of net photosynthetic rate ofginseng. Then on the basis of adaptive neural fuzzy inference method, designed ANFISprediction algorithm and constructed ANFIS prediction model of the net photosynthetic rateof ginseng. The sample tests data of typical test areas validated that the model has highgeneralization ability and prediction accuracy.(4) In view of the problems that exists in traditional data acquisition methods of lightenvironment of ginseng, the paper designed a set of real time monitoring and dataacquisition system. On one hand, using the adaptive data processing algorithm, compiledthe acquisition system underlying algorithm and graphical user interface software,completed the design of software architecture; On the other hand, using the Arduinomicrocontroller open source control platform, completed the hardware platform designincluding the temperature and humidity collecting sensor module based on DHT11of datatemperature and humidity sensor, atmospheric pressure acquisition module based onBMP085pressure sensor, the light intensity acquisition module based on BH1750FVI light sensor and data storage module based on MicroSD card, all which realized theimplementation of real-time monitoring for light environment data of ginseng under forest,at the same time, the system software integrated with the algorithm for adaptive dataprocessing has a well noise reduction function.(5) Based on the analysis of the suitability of geography, climate and vegetationecological conditions for cultivated ginseng under forest, and using fuzzy set theory, thepaper has researched and designed the ecological suitability evaluation index of lightenvironment and constructed the comprehensive evaluation model based on fuzzy inferencesystem including geographical condition evaluation subsystem, climate conditionevaluation subsystem, vegetation condition evaluation subsystem and ecological conditionevaluation of subsystem. And then build the light environment prediction and evaluationsystem of cultivated ginseng under forest. The system integrates the trees growth modelprediction algorithm, as well as the ANFIS model prediction algorithm of netphotosynthetic rate, all of which can, according to the given regional features of cultivatedginseng under forest, predict the models of trees factors, and on the basis of comprehensiveanalysis of the existing data, make comprehensive evaluation for light environmentsuitability of cultivated ginseng under forest.