| The unique advantages of light and heat resources make Xinjiang an important forest and fruit producing area in China.Ziziphus jujuba,as one of the characteristic forest fruit of Xinjiang forest and fruit industry,ranks first in the country in planting scale and output,and is the pillar industry of rural economic development in southern Xinjiang.However,due to the low level of fruit farmers’ cultivation and management techniques,the problem of jujube diseases and insect pests restricts the quality and yield of jujube.Hyperspectral remote sensing,as an accurate,fast,and convenient spectral detection technology,provides a technical way for realtime,rapid and efficient monitoring of plant nutrition,growth,health status,etc.It is of great practical significance to explore the use of hyperspectral technology for monitoring vegetation pests and diseases.In this study,jujube leaves under the damage of Tetranychus truncatus in the planting area of Angtle County,Maigaiti County,Kashgar region were taken as the research objects,the changes of leaves spectra and physiological parameters,the correlation between physiological parameters and spectrum,and the model of estimation of Tetranychus truncatus hazard level were studied.The main results were as follows:The spectral characteristics of the leaves in the different damage period of Tetranychus truncates showed that:The original spectral curve of healthy plant leaves was significantly higher than that of pest damage period,and gradually decreased with the advancing spectral curve of damage period,and the first order differential spectral curve increases with the increase of hazard level,while the first-order differential spectral curves showed a "blue shift" phenomenon in the red edge region.The changes of physiological parameters of jujube leaves were very obvious in the different damage period of Tetranychus truncates.The chlorophyll content and water content of leaves gradually decreased with the increase of the hazard level of Tetranychus truncates.The correlation analysis between jujube leaves and physiological parameters of different damage period of Tetranychus truncates showed that:the correlation between the first-order differential spectra and physiological parameters was significantly higher than that of the original spectra.The absolute value of the maximum correlation coefficient between the original spectrum and chlorophyll content of Tetranychus truncates hazard period is only 0.450,the absolute value of the first order differential maximum correlation coefficient is 0.740,the absolute value of the original spectrum and moisture content maximum correlation coefficient of each Tetranychus truncates damage period is only 0.370,and the absolute value of the first order differential maximum correlation coefficient is 0.780.The sensitive bands of physiological parameters and the spectrum of jujube leaves were mainly concentrated in the near-infrared(700~1 300 nm)and medium-infrared(1 300~2 500 nm)ranges.The study of hyperspectral estimation model of leaves physiological parameters showed that the estimation model of leaves physiological parameters of jujube leaves with multiple linear regression was better than that of single factor regression and principal component regression.A model that can estimate chlorophyll content in jujube leaves during the various damage period of Tetranychus truncatus fitting degree was between 0.460~0.880,and the burst model of multiple linear regression model is best fitted.A model that can estimate water content in jujube leaves during the various damage period of Tetranychus truncatus fitting degree was between 0.61 7~0.914,and the pre-explosion model of multiple linear regression model is best fitted.Study on hazard level model of Tetranychus truncates indicates that:based on the results of physiological parameter change and correlation analysis between spectrum and hazard level of j ujube leaves under the damage of Tetranychus truncates,chlorophyll content,water content,original spectral sensitive band 729 nm、first-order differential spectral sensitive band 642 nm、644 nm、716 nm、717 nm、718 nm and 723 nm were selected as independent variables,and the Fisher discriminant analysis method was used to establish the hazard level detection model of Tetranychus truncates,the total discriminant accuracy is 81.4%,the test discriminant accuracy is 71.5%,and the prediction accuracy is 77.3%,which indicated that the Fisher discriminant analysis method was feasible for detecting the hazard level of Tetranychus truncatus. |