Continuous Wavelet Transform and Back Propagation Neural Network for Condition Monitoring Chlorophyll Fluorescence Parameters Fv/Fm of Rice Leaves

The chlorophyll fluorescence parameter Fv/Fm (maximum photosynthetic efficiency of optical system II) is an intrinsic index for exploring plant photosynthesis. Hyperspectral remote sensing technology can be used rapid nondestructive detection parameters. Existing studies show that there a good correlation between the vegetation and Fv/Fm. However, due to limited hyperspectral information reflected by index, established model often cannot reach ideal accuracy. Therefore, this study took rice as research object explored internal relationship parameters spectral reflectance setting different fertilization treatments. Spectral sensitive was extracted continuous wavelet transform (CWT) explore more suitable method estimation at leaf scale. Then monitoring in leaves back propagation neural network (BPNN) algorithm. results showed that: (1) accuracy univariate models constructed inversion based on 10 commonly indices traditional methods low; (2) could effectively improved using CWT, best coefficients level rough evaluation Fv/Fm; (3) effect mother functions basis function different, bior3.3 best; R2, RMSE RPD BPNN first decomposed 0.823 6, 0.013 2 2.304 3. In conclusion, proves CWT extract bands monitoring, providing reference follow-up fluorescence.