Identification Of Citrus Huanglongbing Based On LIBS-NIR

Huanglongbing(HLB)is the most severe infections caused by gramin-negative bacteria(Candidatus Liberibacter asiaticus,CLas).It seriously affects the growth of trees and the quality of fruits,causing huge losses to the navel orange industry worldwidely.As of now,there is no fundamental and effective solution.Spectrometry obtains the advantages of fast detection speed and no complex pretreatment.It can realize accurate,fast,pollution-free detection.In this paper,the feasibility of rapid detection of HLB by LIBS-NIR was explored with "Gannan navel orange" as the research object.The main research contents are as follows:(1)The differences in sugar content and mineral element content in healthy and HLB-infected navel orange were analyzed,and it revealed the effect of HLB on the internal quality and physicochemical properties of navel orange.It offers an experimental basis for rapid detection of HLB by spectrometry,and a new thought for analysis of HLB based on the element level.The results showed that:(1)There were some differences in sugar content between healthy and HLB-infected navel orange.The average sugar content of healthy navel orange was 11.3 Brix,while that of HLB-infected navel orange was only8.6 Brix.Sig value was 0.000,when the Kolmogorov-Smirnov test(K-S test)was used to investigate the difference between the sugar content in healthy and HLB-infected navel orange,and it indicates that the sugar content of healthy and HLB-infected navel orange is significantly different.(2)The content of mineral elements in HLB-infected navel orange changed to some extent.The average content of Ca in normal navel orange leaves decreased from 15148 mg/kg to 7625 mg/kg.The average content of Ca in navel orange leaves decreased from 15148 mg/kg at normal to 7625 mg/kg,the average content of Zn in leaves decreased from 198 mg/kg at normal to 70 mg/kg,and the average content of Zn in navel orange peel increased from 67 mg/kg at normal to 139 mg/kg.The K-S test showed that the Sig values were all less than 0.05,indicating that the change was significant.(2)Based on Laser induced breakdown spectroscopy(LIBS),the navel orange peel as the research object,the HLB identification model was established.It realizes speedy identification of HLB,and provides technical assistance for intelligent mechanical removal of HLB-infected navel orange fruit and removal of susceptible plants.The LIBS plasma time-evolution morphologies and spectra of healthy and HLB-infected navel orange peel were collected to analyze their differences in plasma morphology and LIBS spectra.Continuous wavelet transform(CWT)and Principal component analysis(PCA)are proposed to reduce the influence of unrelated variables on the test and improve the efficiency of the identification model.The results showed that:(1)The plasma volume of navel orange peel was the largest when the collection delay time was 1.0 ~ 2.0 ?s,and theplasma morphology of healthy and HLB-infected navel orange peel had some differences,which indicated the feasibility of using LIBS technology to quickly detect HLB.(2)Analysis of the major element characteristic line intensity of navel orange peel and other major element values of K element internal standard,the LIBS spectrum of healthy and HLB-infected navel orange navel orange peel has certain difference,and the content of Mg and H in the diseased navel orange peel decrease,while the contents of Ca and Na increase.(3)Using CWT and PCA combined with Fisher discriminant analysis(FDA)and Multilayer perception neural network(MLP)to identify HLB,and the accuracy of HLB is above 94.0%,PCA-MLP training set and prediction set accuracy rate reached 100.0%.(3)Based on Near infrared spectrum(NIR)and LIBS,the navel orange peel as the research object,and the LIBS-NIR technique with greater detection advantages was proposed,which realized the rapid and accurate identification of HLB.It reveals the detection merits of the fusion spectrum compared to the single spectrum,and provides a reference for joint spectrum detection in other fields.The possible reasons for the difference of plasma morphology,LIBS and NIR spectra between healthy and HLB-infected navel orange leaves were analyzed theoretically.Baseline correction was used to reduce the impact of background signal on spectral quality.The results showed that:(1)Analysis of the major element characteristic line intensity of LIBS and the other major element values of K element internal standard showed that the content of Mg,Ca,C,H and O in the leaves of HLB-infected was decreased,the content of Na was increased.,which indicated the feasibility of using spectral detection method to predict HLB.(2)The LIBS-NIR obtained higher discrimination accuracy than the single spectrum.The fusion spectrum combined with the PCA-MLP model training set identification accuracy reached89.5%,and the prediction set accuracy reached 95.7%.