Study On Identification Method Of Tea Algal Spot Disease Stage Based On Fluorescence Spectrum

Tea leaf was one of the most main drinks in China,which was one of the country that the tea plant scale was largest,while the various diseases have been seriously affecting the yield and quality of tea.Tea algae leaf spot as one of the main diseases,can stress the normal metabolism of tea tree,decrease of tea yield and quality,and cause direct economic losses to tea growers,covers a vast area and occurs in the main tea areas of China.At present,the traditional detection methods for algal spot disease are difficult to be widely applied in practical production due to their long time limit,complicated pretreatment and high cost,therefore,the key to the development of tea industry is to accurately identify tea diseases and take timely preventive measures.In this study,the identification research of tea-algal blotch based on fluorescence spectrum was carried out to explore a nondestructive and rapid detection method of tea algal spot,and to provide a reference for the application of chlorophyll fluorescence combined with remote sensing technology in large area of agriculture.The main research contents and conclusions are as follows:(1)The chlorophyll content of normal,slight and serious algal-spot leaves was measured.One-way ANOVA was performed on the chlorophyll content of the three kinds of leaves.The results indicated that there were significant differences between the three kinds of leaves,but the differences among the three kinds of leaves were not explained.Then,multiple correlation analysis was conducted for chlorophyll content values of the three types of leaves.The results of Tamhane’s T2 and Dunnett’s T3 demonstrated that the probability of pair significance value among the three types of leaves was less than 0.05,indicated among three kind of leaves has significant differences.The results lay a theoretical foundation for the identification of tea algal spot disease based on chlorophyll fluorescence spectrum.(2)Normal,slight and serious algal-spot leaves as the research object,and theirs differences of chlorophyll fluorescence spectra were analyzed.Principal component analysis(PCA)and successive projection algorithm(SPA)were used to select spectral variables.Full spectral variables and those selected by PCA and SPA were used as input of support vector machine(SVM)and partial least squares discriminant analysis(PLS-DA)models,and SVM model and PLS-DA model with different kernel functions were established.It showed that the variable selection algorithm was not suitable for the PLS-DA model,but has a better effect on the SVM model,and has the best effect in the RBF-kernel model.SPA-RBF-SVM model was better in the discrimination of 3 kinds leaves,and the misdiagnosis rate was 4.69%.(3)The substances have different selectiveness to light sources,and the fluorescence efficiency by different wavelengths excited sources varies greatly.The leaf chlorophyll fluorescence excited by Light Diode(LD)laser with a central wavelength of 405 nm and Light Emitting Diode(LED)with a central wavelength of 405,470,533 and 635 nm were collected and the spectral differences were analyzed.A quantitative analysis model of chlorophyll content in normal slight and serious algal spot disease mixed leaves with different light sources was established to study the influence of different light sources on the quantitative model of chlorophyll content in three kinds of mixed samples.The results concluded that LED 405 nm as the excitation light source of chlorophyll fluorescence has achieved better effect,and the correlation coefficient and root mean square errors were 0.90,6.10 and 0.88,6.10 in set calibration set and validation set respectively.(4)The reabsorption effect of chlorophyll fluorescence was changed with the fluorescence path of leaves changed in different collecting angles.Based on the study of the different light source model,used 405 nm LED excitation light source as a benchmark,designed the 90°,75°,60° and 30° of acquisition experiments,collected different angles of chlorophyll fluorescence,the quantitative analysis models of fluorescence spectra and chlorophyll content of three kinds of mixed samples with different collection angles were established to study the effect of different collection angles on the model.The results pointed that when the acquisition angle is90°(i.e.,the optical fiber probe is collected vertically with the horizontal blade),the model has achieved satisfying effect.(5)On the basis of the best light source and the best acquisition angle,the effects of different preprocessing methods and variable optimization algorithms on the model effect were studied,and the results showed that the preprocessing of the first derivative significantly improved the model.Secondly,four variable screening algorithms were combined to optimize the variables,and the best quantitative prediction model of algal spot disease was established.The prediction performance of the model was evaluated for normal,slight and serious algal spot disease leaves and three mixed samples.Correlation coefficient and root mean square prediction error of prediction was 0.91,5.88;0.74,7.66;0.86,6.81 for normal,slight and serious algal spot disease leaves respectively.And prediction correlation coefficient of three kinds of vane mixed samples and root mean square prediction error was 0.87,6.84.