应用遗传算法结合连续投影算法近红外光谱检测土壤有机质研究

应用遗传算法结合连续投影算法近红外光谱检测土壤有机质研究。采集浙江省文城地区农田土壤样品近红外光谱数据,土壤样品数为394个。为简化模型,采用遗传算法结合连续投影算法挑选出18个特征波长建模,应用偏最小二乘回归建立有机质预测模型,建模集的决定系数为0.81,均方根预测误差为0.22,剩余预测偏差为2.31,预测集的决定系数为0.83,均方根预测误差为0.20,剩余预测偏差为2.45。研究发现,遗传算法结合连续投影算法在简化模型同时,模型的预测评价指标同采用全谱波长建模并没有明显降低。因此,遗传算法结合连续投影算法挑选的特征波长可以应用于近红外光谱检测土壤有机质含量。

Measurement of Soil Organic Matter with Near Infrared Spectroscopy Combined with Genetic Algorithm and Successive Projection Algorithm

Visible near infrared spectroscopy combined with genetic algorithm and successive projection algorithm was investigated to detec t soil organic matter (OM).A total of 394 soil samples were collected from Wenc heng,Zhejiang province.In order to simplify calibration model,a total of 18 c haracteristic wavelengths were selected with usinggenetic algorithm and successi ve projections algorithm.These characteristic wavelengths were subjected to par tial least squares regression (PLSR) with leave-one-out cross validation to es tablish calibration model of soil organic matter (OM) with coefficient of determ ination (R2) of 0.81,0.83,RMSEP of 0.22,0.20 and residual predictio n deviation (RPD) of 2.31,2.45 for the calibration set and prediction set res pectively.The results showed that using genetic algorithm and successive projec tions algorithm can simplify the model greatly while the assessing indexes of mo del such as R2,RMSEP and RPD were not reduced greatly compared with index es of model using full spectra data to develop calibration model.Therefore,gen etic algorithm combined with successive projections algorithm can be used to sim ply the model to predict soil organic matter.