Measurement Of Key Parameters In Agriculture And Fruit Farming Based On Spectroscopy

Precision Agriculture is a symbol of modern agricultural technology based on information and knowledge.Because spectral technology has the characteristics of real time measurement,convenience and environmental protection,it can meet the requirements of modern agricultural information acquisition.In this paper,soil and apple\citrus tree were regarded as the research objects.The main contents and results of this paper are as follo,ws:[1]The distribution of soil nitrogen in longitudinal space was investigated.Then the correlation coefficient was calculated between the first order differential spectral information and the total nitrogen content.Based on the correlation analysis result,the characteristic wavelengths were selected for different soil layers,and then the multipal prediction models of soil total nitrogen were established.Finally,the BP neural network model optimized by genetic algorithm was used to retrieve the total nitrogen content of soil in different layers.[2]The original soil spectral information was denoised using wavelet filtering technique.And the relative SNR parameter was constructed to evaluate the wavelet denoising.The envelope-removing processing was done among the optimal filtering signals.Then 6 wavebands(1375,1520,1861 2100,2286,and 2387nm)which could reflect the soil total nitrogen content under 5 different moisture gradient conditions were got.Therefore,in this study,under the condition of no manual pretreatment,like drying and sieving,the spectral information of nitrogen in the soil was effectively extracted from the original spectra of the soil.[3]A new criterion used to evaluat the sensitive wavelength of soil total nitrogen content was proposed.The evaluation criteria and ant colony optimization algorithm were combined to make full use of the advantages of information theory and ant colony algorithm in feature extraction.And soil samples after variable rate fertilization with 3 different types of nitrogen fertilizer(urea,calcium nitrate,potassium nitrate)were used to verify the selected sensitive wavelength.The results show that the selected wavelengths have strong robustness in the detection of soil nitrogen.[4]A non-contact system for measuring soil moisture and total nitrogen content based on laser light source was developed.The octagonal "umbrella" structure was applied as the core part of the optical unit to realize the optical signal transmission and acquisition.In order to meet the requirements of field measurement with wide range and long working time,the wireless communication mode was adopted to solve the problem of real-time data acquisition and transmission.The experimental results show that the device has good accuracy,stability and detection ability of soil total nitrogen and moisture content.[5]After field experiments in several years,the VIS/NIR spectral signature of apple tree leaves in some important phonological phases including fruit-bearing period,fruit-abscission period,fruit-mature period was investigated.The normalized differential vegetation index(NDVI)focused on the nitrogen content of apple tree leaves was built.Then wavelet and wavelet packet technologies were used to extract the effective spectral information which could reflect nutrient status of apple tree leaves.Many nutrition prediction models were established later.These models using the selected wavebands showed good veracity in predicting nutrient elements content of apple tree leaves in different periods,which can be used as the best estimation model in this phase.[6]Two-dimensional correlation analysis was conducted between fruit sugar content with the leaves spectral information in 6 phenological phases including flowering period(F.P.),shoot-growing stage(S.G.S.),fruit setting period(F.S.P.),branch shooting period(B.S.P.),bud differentiation stage(B.D.S.)and defoliation period(D.P.).The sensitive wavebands of leaves reflecting fruit sugar content were obtained.Through the principal component regression,the changes of photosynthesis intensity and the contribution weights to the sugar content in different phases were investigated.Then the model based on parametric optimal solution of SVM showed good accuracy,which testified the validity of fruit sugar prediction using spectral information of leaves.[7]The citrus diseases were detected based on spectroscopy in order to propose a universal algorithm for extracting the spectral features of citrus fruits.The successive projection algorithm(SPA)and ant colony optimization(ACO)algorithm were combined with wavebands selection principal based on correlation coefficient separately,which was applied in spectral signature selection of typical citrus diseases.Then the 'simple-layer' and 'tree-type' classification models based on SVM were designed.After that,the healthy fruit and the diseased fruit with 6 typical diseases(HLB,melanoses,oil spot,wind scar,insect damage.rust mites)were successfully classified with high accuracy.It lays an important foundation for the development of Citrus disease identification system.In this paper,the visible and near infrared spectroscopy technology was used to realize the real-time monitoring of important parameters of soil and fruit trees.The study provides a reliable scientific basis for the implementation of spectral nondestructive detection under various agricultural applications,and provides a strong technical support for the further realization of precision agriculture.