Detection Of Antioxidant Enzyme Activity In Tomato Leaves Based On Microscopic Hyperspectral Imaging Technology

Solarium lycopersicum plays an important role in people’s life and is one of the indispensable vegetables.However,it faces the problem of salt stress in the cultivation process,which inhibits the normal growth and development of tomato and reduces the fruit quality of tomato.The antioxidant enzymes in plants are important parameters to reflect the antioxidant capacity of plants.This test set based on microscopic highlights like technology for salt stress under the condition of POD,CAT and SOD three chemical indicators for testing,using chemical index,spectral information fusion of hyperspectral image,build salt stress under the condition of tomato leaf antioxidant enzymes macroscopic quantitative model,realize single antioxidant enzyme spectrum in the cell detachment and visual distribution.The interaction of antioxidant enzymes in tomato leaf cells was clarified,and the high-precision in-situ rapid detection technology system of antioxidant enzymes in tomato leaf cells was constructed,providing theoretical reference for the rapid detection technology development of other microscopic substances in tomato leaf cells.The main research contents and results are as follows:(1)RS method was used to divide 298 tomato leaf samples,and six pretreatment methods were used to optimize the spectrum.The results showed that POD activity of PLSR model based on Baseline was improved(Rc=0.8203,Rp=0.8010,RMSEC=90.4396,RMSEP=107.1999).CAT activity was best modeled by original spectrum(Rc=0.8609,RMSEC=0.7739,Rp=46.5239,RMSEP=62.8269).Normalize method was the most effective method for SOD optimization(Rc=0.8888,Rp=0.8937,RMSEC=36.1952,RMSEP=41.0579).(2)The prediction models of POD activity,CAT activity and SOD activity in tomato leaves were established based on the optimized characteristic wavelengths.Cars-plsr was used as the best prediction model for POD activity,and the correlation coefficient of the corrected set was Rc=0.8714 and the correlation coefficient of the predicted set was Rp=0.8664.The optimal prediction model of CAT activity was IRF-PLSR,Rc=0.8749,Rp=0.8086.The effect of CARS-LSSVM model on SOD activity was relatively good,Rc=0.9427,Rp=0.9058.The results showed that hyperspectral imaging technology could be used to predict the activity of antioxidant enzymes in tomato leaves,which provided a scientific theoretical basis for the effective detection of their distribution and rational management.(3)Combined with the macro established model of antioxidant enzyme content in tomato leaves,according to the ratio of the microarea to the slice area of macro tomato leaves,the migration calculation of antioxidant enzyme activity in the observation area of macro tomato leaves and the activity of antioxidant enzyme in the microarea was carried out to establish the prediction model for cell image and spectral detection in the microarea.The predicted POD activity was Rc=0.8528,Rp=0.8323.The predicted CAT activity was Rc=0.7896,Rp=0.7146.The predicted results of SOD activity were Rc=0.9751,Rp=0.9368.(4)Visual study was carried out by combining three chemical indexes of POD,CAT and SODin tomato leaves with microscopic hyperspectral data.The prediction model based on optimal pretreatment was established according to the chemical index.The β coefficient method was used to extract the spectral input variables and obtain the visual imaging effect map,which provided a reference for the in-depth study of spectral technology in plant microarea detection.