Study On Chlorophyll And Nitrogen Content Of Walnut Forest Crown Based On Hyperspectral Nnformatio

Chlorophyll and nitrogen are important physiological and ecological indicators of plants.Understanding the chlorophyll and nitrogen content of walnut forest canopy at different growth stages has important guiding significance for guiding the scientific fertilization and precise management of walnut forests.In this study,the 5-year-old Xiangling walnut tree was used as the research object,and the hyperspectral information of walnut forest canopy in walnut fruit expansion period,hard-core period and oil conversion period was obtained using an airborne hyperspectral imaging system.Based on the accurate extraction of the effective information of the walnut tree canopy,the vegetation index method（Vegetation Index,VI）screens the characteristic sensitive bands indicating the chlorophyll and nitrogen content of the walnut canopy at different growth stages,and obtains the reconstructed vegetation index based on the characteristic sensitive bands.The calculation models of canopy chlorophyll and nitrogen content in walnut fruit expansion period,fruit hard-core period and oil conversion period,and the calculation models established in each growth period are applied to walnut forest land of different forest ages and annual time series for model verification.Formulating refined water and fertilizer management,accurately assessing growth and predicting output are important guidelines.The main findings are as follows:1.Spatial variation characteristics of chlorophyll and nitrogen contentThe chlorophyll and nitrogen content of walnut tree canopy in the four directions east,west,south,and north and the three sample trees fluctuate little,and the SD dispersion is low.The chlorophyll and nitrogen content between the four directions and the three sample trees The content showed no significant difference by analysis of variance.Therefore,in actual measurement,a large number of chlorophyll and nitrogen contents in 4 directions can be measured,and the average value is used to characterize the chlorophyll and nitrogen status of the entire canopy.2.Screening study on accurate extraction methods of effective information within the canopy of walnut treesIn this study,six remote sensing image classifiers including parallelepiped,minimum distance,Mahalanobis distance,maximum likelihood,neural network and support vector machine were used to initially extract the range of walnut tree canopy.The preliminary extraction results were subjected to Majority,filtering and clustering spots.After the block is removed,comparative analysis is performed.Among them,the support vector machine classifier has the highest recognition accuracy of the walnut tree canopy after Majority patch processing.Its overall classification accuracy is 100,Kappa coefficient is 1.0,misclassification error is 0.00%,and omission error is 0.00%.The drawing accuracy is 100%,and the user accuracy is 100%.At the same time,based on the spectral feature technology to find the band interval that does not overlap with other non-target vegetation,the target vegetation walnut and other non-target values are effectively identified,and the support vector machine method is used to verify the accuracy.The classification accuracy is as high as 96.43%.3.Establishment and verification of the estimation model of chlorophyll content in walnut canopyBased on the correlation between the chlorophyll and the original spectral reflectance,the single band B192（707.4）was determined as the characteristic sensitive band indicating the chlorophyll content of walnut fruit during the swelling period and the fruit hard-core period;the single band B196（714.9）was the chlorophyll indicating the walnut oil conversion period The characteristic sensitive band of content.According to the correlation between the two-band combination of chlorophyll and spectral reflectance,it is determined that the two-bands B237（786.5）and B200（721.4）,B247（804.3）and B193（709.2）,B233（779.4）and B202（724.9）respectively indicate walnut fruit The characteristic sensitive bands of chlorophyll content in the canopy of the swelling stage,the fruit hard-core stage and the oil conversion stage.According to the correlation between the chlorophyll content and the three-band combination,the three-band B69（499.3）,B200（721.4）,B237（786.5）,B10（403.7）,B193（709.2）,B265（836.5）and B21（421.3）,B197（716.1）and B319（934.6）respectively indicate the characteristic sensitive bands of canopy nitrogen content in walnut fruit swelling stage,fruit hard-core stage and oil conversion stage.The reconstructed vegetation index was constructed based on the characteristic sensitive bands screened in each growth period to establish the estimation model of the chlorophyll content of the walnut canopy,and the calculation model after leaving one cross-validation was applied to the similar-age walnut forest for verification.The results show that the calculation model established by the reconstructed vegetation index EVI obtained by the three-band combination in each growth period of the walnut tree has the highest accuracy,and the calculated value is closer to the measured value,which has higher model stability and reliability.4.Establishment and verification of the calculation model of walnut canopy nitrogen contentBased on the correlation between nitrogen and the original spectral reflectance,the single bands B196（714.9）,B192（707.4）and B193（709.2）were determined to indicate the canopy nitrogen content of walnut fruit swelling stage,fruit hard core stage and oil conversion stage,respectively.According to the correlation between the nitrogen content and the two-band combination of spectral reflectance,the two bands B347（986.4）and B186（697.0）were identified as characteristic sensitive bands indicating the canopy nitrogen content in walnut fruit swelling stage and fruit hard-core stage;The bands B183（691.8）and B32（439.0）indicate the characteristic sensitive bands of the nitrogen content in the canopy during the oil conversion period of walnut trees.Construct a reconstructed vegetation index based on the characteristic sensitive bands screened in each growth period to establish a measurement model for the nitrogen content of the walnut canopy,and compare the calculation capabilities of the six published spectral parameters that can be used for the nitrogen inversion of commercial forest tree species in this study.It was applied to walnut forests of similar forest age for verification.The results showed that:the estimation model of canopy nitrogen content in walnut tree fruit swelling stage established based on the reconstructed vegetation index NDVI（986.4,697.0）has high accuracy,the maximum verification R² is 0.784,the verification RMSE and verification RE are less than10%,9.784%and 2.938%respectively;The measurement model of canopy nitrogen content in walnut tree fruit at the hard-core stage based on the reconstructed vegetation index NDVI（986.4,697.0）has a high accuracy.The maximum verified R² is 0.746,and the RMSE and verification RE are 10.463%3.127%respectively;The power model y=5.112x^0.623 established based on the reconstructed vegetation index NDVI（691.8,439.0）is a model for measuring the nitrogen content in the canopy of walnut trees during the oil conversion period.The verification R² is 0.717,which verifies the RMSE And the verification is 12.114%,4.325%,respectively,the model estimation accuracy is higher.