Study On Fourier Transform Infrared Spectroscopy Of Citrus Disease

The quantity and quality of fruit are affected by diseases during growing and afterharvest, and the benefits of farmers are also influenced by the diseases. It is benificial toimplement continous devolopment through rapid diagnosis and discrimination ofcitrus dizeases. Fourier transform infrared (FTIR) spectroscopy was usd to studyseven common citrus diseases.Citrus brown spot, huang long bing, canker, fuliginous, cercospora sp. andhealthy leaves were studied using FTIR spectroscopy combined with statisticalanalysis. There were obvious differences in the range of1200~700cm-1of secondderivative spectra, the performance of the overall accuracy of principal componentanalysis (PCA) based on second derivative dataset was92.5%.Partial Least Squares-discriminant analysis (PLS-DA) and Back PropagationNeural Network (BPNN) based on Continuous wavelet transform (CWT) is applied toidentify citrus tatter leaf (CTL) and healthy leaves by FTIR spectroscopy. CWT isapplied to analyze the FTIR spectra of60samples by selecting Haar, Mexh andMorlet wavelet. By comparison, the decomposition level7is obvious differences, andthree regions of this level are selected as feature vectors. This feature vectors are usedto train PLS-DA and BPNN models. The accuracy of BPNN (95%) is better thanPLS-DA (75%).Linear Discriminant Analysis (LDA) and BPNN based on wavelet transform(WT) is applied to identify Citrus sinesis (L.) Osbeck anthracnose and healthy peel byFTIR spectroscopy. Discrete wavelet transform (DWT) is used to compose spectra of60samples in1750-950cm-1range. Discrete wavelet transform approximationcoefficients (DWTAC) and discrete wavelet transform detail coefficients (DWTDC)of level5are classified using LDA and BPNN. Results show that the accurate recognition rate by LDA and BPNN models based on DWTDC is95%.It shows that FTIR spectroscopy combined with chemometrics can be used foridentification different citrus diseazes accurately, It also provides technology supportto detect citrus deseases in early stage quickly and effectively.