It was successful to de-noise the spectrum signalwithin visual wave band (350 550nm) by wavelettransformation and extract some characteristic wave-peaks such as chlorophyll's 380nm , 414nm , 437nmand the folic acid 366nm. In the range from 550nm to2500nm, after de-noising, the biggest error was smallerthan 1.47%, while at the peaks of waves the biggest errorwas smaller than 0.11%. The results showed that thewavelet de-noising used in this work slightly or not affectthe other non-noise band. The method is very effective inpreprocessing spectral data. More consideration of thechlorophyll spectrum characters is helpful to provide abasis for developing chlorophyll meter which has aStrong anti-interference, high reliability and universality.
展开▼
