【摘要】 为了准确监测辣椒生长，本研究对辣椒冠层光谱反射率进行对数处理、倒数处理、倒数的对数处理、连续统去除处理、一阶微分处理、二阶微分处理，并与SPAD值进行相关性分析，用最大相关系数法(MCC)选取相关性较好的特征波段生成特征波段数据集，再用遗传算法-偏最小二乘法(GAPLS)进行降维得到最优特征波段组合，采用偏最小二乘法(PLSR)、反向传播神经网络(BPNN)、随机森林(RF)和最小二乘支持向量机(LSSVM)4种机器学习算法构建辣椒叶绿素含量反演模型。结果表明，最优波段和对应处理分别为700 nm(原始光谱)、699 nm(对数处理)、713 nm(连续统去除处理)、500 nm(二阶微分处理)、713 nm(二阶微分处理)。GAPLS的降维效果较好，与降维前相比PLSR模型的精度提升率最高，R~2、RPD分别提升了82.22%、136.98%,RMSE降低了29.96%。4种模型中，GAPLS降维处理后的PLSR模型的精度最好，R~2、RMSE和RPD分别为0.82、1.94、4.55。本研究构建的MCC-GAPLS-PLSR模型具有较好的反演潜力，适用于研究区辣椒叶片叶绿素含量测定，推动辣椒高效种植。更多还原

【Abstract】 In order to accurately monitor the growth of peppers, this study performed logarithmic treatment, reciprocal treatment, reciprocal logarithmic treatment, continuum removal treatment, first derivative treatment, second derivate treatment on the canopy spectral reflectance of peppers, and conducted correlation analysis with SPAD values. The maximum correlation coefficient method(MCC) was used to select the feature bands with good correlation to generate a feature band dataset. And the genetic algorithm-partial least squares(GAPLS) was used to reduce the dimensionality to obtain the optimal feature band combination. Pepper chlorophyll content inversion model was constructed by using four machine learning algorithms: partial least squares regression(PLSR), backpropagation neural network(BPNN), random forest(RF) and least squares support vector machine(LSSVM). The results showed that the optimal wavelengths and corresponding treatments were 700 nm(original reflectivity), 699 nm(logarithmic treatment), 713 nm(continuum removal treatment), 500 nm(second derivate treatment), 713 nm(second derivate treatment). The dimensionality reduction effect by GAPLS was good. And compared with before dimension reduction, the accuracy improvement rate of PLSR model was the highest, R~2 and RPD increased by 82.22% and 136.98% respectively, and RMSE decreased by 29.96%. Among the four models, PLSR model after dimensionality reduction by GAPLS had the best accuracy, with R~2, RMSE, and RPD of 0.82, 1.94, and 4.55, respectively. The MCC-GAPLS-PLSR model constructed in this study has good inversion potential and is suitable for measuring the chlorophyll content of pepper leaves in the study area, thus promoting efficient cultivation of peppers.更多还原