PLS算法在高光谱估测加工番茄白粉病色素含量中的应用!

PLS Algorithm Application in Hyperspectral Estimation of Pigment Contents in Processing Tomato Leaves under Powdery Mildew

采用传统光谱特征参数和新建光谱特征参数对色素含量进行多光谱特征参数的偏最小二乘法(PLS)估算,以此来估测加工番茄白粉病病叶色素含量,并进行精度检验。结果表明:色素含量原始光谱与反对数光谱的敏感波长分别为350～718 nm、719～839 nm和350～718 nm、728～857 nm。从传统光谱特征参数和新建光谱特征参数中选取9个光谱特征参数PSNDa、GNDVI、PSSRa、PSSRb、GM、TPMPa、TPMPb、TPMPc和TPMPd,分别构成PGP、PGT、TTT 3组对色素含量进行估测,TTT组为色素含量光谱组合最佳估测模型。

Both traditional and new spectral characteristic parameters were used to estimate the pigment content of leaves by employing PLS algorithm.Therefore the pigment content of leaves under powdery mildew in processing tomato could be estimated and tested.The result showed that the sensitive bands of original and opposing spectrum were 350～718nm,719～839nm and 350～718nm,728～857nm,respectively.Then three groups named PGP,PGT and TTT were composed,including the optimal 9 spectral characteristic parameters from either traditional or new parameters（PSNDa,GNDVI,PSSRa,PSSRb,GM,TPMPa,TPMPb,TPMPc and TPMPd）.After estimating the pigment content,the best estimation model was TTT.