基于离散小波算法定量反演贵州 百香果叶片叶绿素含量的研究.

[Objective] The objective of the paper is to explore a rapid and accurate method for diagnosing the biochemical components of Guizhou premium fruit plant. [ Method] Taking hyperspectral data of Passiflora edulis leaves and its corresponding leaf chlorophyll content as the data source, mathematical transform and discrete wavelet transform algorithm were used to process and analyze the spectrum data of P.edulis leaves, the sensitive bands were extracted by correlation analysis, and then the estimating model of chlorophyll content in P.edulis leaves was established by using partial least square algorithm. [ Result] The low frequency information of logarithmic transform separated from discrete wavelet transform and mathematical transform could significantly improve the estimation of spectrum to chlorophyll content in P. edulis leaves. The model established based on discrete wavelet transform L1 scale was optimum, which had high estimated accuracy and stability with R² = 0. 578 and RMSE = 1. 104. [Conclusion] As affected by complex geographical environment and climate condition, the spectrum data of P. edulis leaves contains a large amount of noise. The high frequency information extracted by mathematical transform and discrete wavelet transform has weak abilitying estimating chlorophyll content in P. edulis leaves. Compared with high frequency information, low frequency information has strong estimating ability of chlorophyll content in P. edulis leaves. [ABSTRACT FROM AUTHOR]