水稻叶片SPAD值高光谱成像估测

Estimation of SPAD value of rice leaves based on hyperspectral image

以水稻叶片为研究对象,基于健康和稻瘟病叶片高光谱图像,运用高光谱特征参数、竞争性自适应重加权(CARS)和主成分分析(PCA)算法选取特征变量,采用偏最小二乘回归(PLSR)、支持向量机(SVM)和反向传播神经网络(BPNN)算法,构建水稻叶片SPAD值高光谱估测模型,并对比分析。结果表明,所有模型均可预测SPAD值,最优模型为PCA-BPNN,其预测集决定系数、均方根误差、相对误差分别为0.8082、2.0783、4.18%。研究表明基于健康和稻瘟病叶片的高光谱图像估测叶绿素含量可行,为水稻健康状况监测、病害影响评估提供理论基础。

Taking rice leaves as the research object,hyperspectral images of rice blast leaves and healthy rice leaves were collected.Competitive adaptive reweighting sampling(CARS)and principal component analysis(PCA)algorithms were used to extract the feature variables of the hyperspectral imagery data.Based on the extracted results and the hyperspectral characteristic parameters,nine hyperspectral estimation models of the SPAD values were constructed by partial least square regression(PLSR),support vector machine(SVM)and back propagation neural network(BPNN)algorithms,and the results were compared and analyzed.The results showed that all the models were able to predict the SPAD values,the PCA-BPNN model was proved to be the optimal model,and the determination coefficients,root mean square error and relative error of prediction set were 0.8082,2.0783 and 4.18%,respectively.Therefore,it was feasible to estimate chlorophyll content based on hyperspectral images of health and blast leaves.The results could provide a theoretical basis for monitoring the health status of rice and assessing the impact of diseases.