The Hyperspectral Characteristics And Inversion Of Potassium Content In Beet Leaves Under Drip Irrigation

?Purpose?By analyzing the correlation between sugarbeet total potassium content and canopy hyperspectral data and agronomic parameters under different potassium levels,the inversion model of canopy spectral reflectance+sugarbeet total potassium content and agronomic parameters=sugarbeet plant nutrition parameters was constructed to realize the hyperspectral monitoring of total potassium content in sugarbeet organs and provide support for rapid non-destructive diagnosis of sugarbeet potassium in the field.?Method?KWS9147?Beta vulgaris L.?was used as the test material to set up four field control experiments of potassium application levels K0?0 kg/hm²?,K1?67.5 kg/hm²?,K2?135kg/hm²?and K3?202.5 kg/hm²?.ASD field hyperspectral spectrometer was used to measure sugar beet canopy reflectance spectral data and total potassium content in various organs.The correlation between the original spectral reflectance,first derivative spectral reflectance and total potassium content in sugar beet canopy was analyzed,and the estimation model of sugar beet hyperspectral data and total potassium content was established.?Result??1?Under the control of potassium fertilizer,the dry matter accumulation of sugarbeet under drip irrigation was significantly different,in which the dry matter accumulation in the aboveground part of K3 treatment was significantly higher than that of other treatments,and the duration of the rapid dry matter accumulation period?26d?and the maximum dry matter accumulation rate(2.75g·plant^-1·d^-1)were the highest.The dry matter accumulation in underground part of K2 treatment was significantly higher than that of other treatments,and the duration of rapid dry matter accumulation in underground part?36 days?and the maximum dry matter accumulation rate(4.10g·plant^-1·d^-1)were the highest.With the advance of the growth stage,the leaf area index??LAI??of each treatment increased at first and then decreased,and reached the peak at the root expansion stage?4.54-5.75?,but there was no significant difference among the treatments.?2?The potassium content in leaves,petioles and roots of sugarbeet increased at first and then decreased with the advance of growth stage,and reached the maximum in the stage of root expansion.The potassium content in leaves and petioles reached the peak in K2 treatment?4.71g/kg and 5.21g/kg,respectively?,and the potassium content in root reached the peak in K3treatment?2.77g/kg?.The yield of sugarbeet was K0<K1<K3<K2,the sugar content was K0<K3<K2<K1,and the sugar yield was K0<K1<K3<K2.There was no significant difference among the three treatments.?3?The"blue shift"phenomenon occurred in the red edge position at different growth stages,and shifted 5nm,7nm and 6nm to the left from the front to the later stage compared with the leaf clump stage respectively.The canopy hyperspectral reflectance at different potassium levels is significantly different in the near infrared band.Compared with K0,the canopy hyperspectral reflectance of K1 and K2 increased by 2.2%-6.3%and 6.3%-11.6%,respectively.The canopy hyperspectral reflectance of K3 decreased by 7.5%-13.1%.?4?The maximum correlation between the total potassium content of beet leaves,petioles and root tubers and the original spectral reflectance is 0.646,0.480 and 0.401 respectively in the visible band,0.684,0.472 and 0.387 respectively in the near infrared band,and the maximum correlation with the first derivative reflectance is 0.720,0.657 and 0.441 respectively in the visible band and 0.699,0.662 and 0.403 respectively in the near infrared band.?Conclusion?The power function y=14.72x^1.391 based on hyperspectral characteristic parameter Dr436 and the univariate linear function y=12.85x+5.673 based on vegetation index Dr725-Dr689 can better predict the total potassium content of beet leaves,R² are 0.720 and0.799 respectively.The binomial function y=-1.946x²+55.61x+0.6155 based on hyperspectral characteristic parameter Dr1168 and vegetation index Dr748/Dr556 can well predict the total potassium content of beet petiole,R² are 0.677 and 0.697 respectively.The established functions are difficult to accurately estimate the total potassium content of root tuber.