Inversion Models For Nuts Quality Of Juglans Regia ‘Xinxin2' Using Leaf Spectral Reflectance

Juglans regia ‘Xinxin2' is one of the main walunt producting area of Nanjiang Basin.The conditions of production and quality were the main factor affected the economic value,the traditional method for measured fruit quality had the hysteresis. So we should explor a method to know the quality of nuts before harvest. Therefore, J.regia ‘Xinxin2' in flourishing period was taken as research object, Using the field living leaf specturm determination combined with the indoor chemical analysis of nutrient elements. Based on the analysis about leaves spectral reflectance in different fruit growth periods and the content of leaves nitrogen(N), phosphorus(P), potassium(K). Obained the spectral sensitive bands, established the relationship between the spectral value of the leaves and the nuts quality, constructed the nuts quality inversion model of J.regia ‘Xinxin2'. The main research result as follows.The leaf reflectance spectral had different spectral characterstics in differernt bands,reflectance of fruit different growth period changed similarly, different bends were slightly different. Reflectance presented zigzag at 310-350 nm, “blue valley” were presented at400-490 nm, “green peak” were presented at 490-610 nm, reflectance decreased “red valley”at 610-680 nm, reflectance increased rapidly formed “red ege” at 680-750 nm, near infrared platform appeared at 750-1 080 nm. There were “two peaks” phenomena for the red edge, this phenomenon was significantly related with growth period, it became obvious with growth development, red shift was presented and the red edge amplitude decreased.The fruit growth periods of built nuts weight inversion models could based on the N, P,K contents sensitive bands of leaves in different fruit growth periods. Used the spectral reflectance or its derivative variable of selected fruit growth periods to inversion nuts weight.Through the sensitive bands in the three periods of fruit-set period, fruit hardcore period and fruit fat changes period to build inversion model had the hightest precision. The nut weight inversion model was ?6581.859362.461125.72xxy(10)-(28)(R2 = 0.9185), the differences between the measured and predieted values was not significant after F test.The nut kernel fat content inversion models had differences due to selected different growth periods.Through the spectral reflectance of sensitive bands in five growth periods to build the nut kernel fat content inversion models had the highest fitting degree. The inversion model was?4310.11734.00075.02xxy(10)-(28)(R2 = 0.9212), the differences between the measured and predieted values was not significant after F test.Used the leaf spectral reflectance or its derivative variable of selected fruit growth periods to estimate nuts nut kernel protein content. Through the sensitive bands in the four periods of fruit-set period, fruit fast-growing period, fruit fat changes period and fruit near mature period to build inversion model had the hightest fitting degree. The inversion model was?2827.00995.00128.02xxy(10)-(28)(R2 = 0.9060), the differences between the measured and predieted values was not significant.The results of the study showed that there was a possibility to estimate nuts quality using leaf spectral reflectance in the early.