农作物生长的胁迫因素光谱甄别模型研究

玉米作为我国重要的粮食产物之一，其生长期间的健康检测一直是农业生产的重要问题。本文以受不同因素影响下生长的玉米叶片为研究对象，采用ASD光谱仪进行叶片光谱采集；对原始光谱数据进行导数（Derivative， D）处理，针对经过求导后光谱部分数据无限趋向0的现象，引入压缩感知（Compressed Sensing, CS）方法，并采用迭代重加权最小二乘（Iterative Re-weighted Least Squares, IRLS）数据重建的方法对光谱数据进行恢复；然后选取竞争性自适应重加权算法（Competitive Adapative Reweighted Sampling, CARS），结合不同试验下的影响因素作为标签提取光谱特征；最后通过多层感知机分类模型（Multi-layer Perceptron, MLP），以达到判别生长状态不佳的农作物所受影响因素的目的。本试验生成的D-CS-CARS-MLP模型的精度相较于传统模型精度有所提高，可以高达99%以上，可以看出该模型可以针对农作物生长状态不佳所受的影响因素进行判别。经过验证，D-CS-CARS-MLP模型具有较好的稳定性和精度，为植被健康生长监测提供了新的思路与方法。

Study on spectral identification model of crop growth stress factors

As one of the important food products in our country, the health detection of maize during its growing period has been an important problem in agricultural production. In this paper, the leaves of maize grown under the influence of different factors were taken as the research object, and the ASD spectrometer was used to collect the spectra of the leaves. The derivative (D) of the original spectral data is processed, and the compressed sensing (CS) is introduced to solve the phenomenon that the spectral data after derivative approach to 0 infinitely, the iterative re-weighted least squares (IRLS) data reconstruction method is used to restore the spectral data. Then competitive adaptive re-weighted sampling (CARS) was used to extract the spectral features, and the multi-layer perceptron (MLP) was used to extract the spectral features, in order to identify the factors affecting the poor growth of crops. The accuracy of the D-CS-CARS-MLP model generated in this experiment can be as high as 99% , and the model can be used to identify a variety of factors. After verification, the D-CS-CARS-MLP model has good stability and precision, which provides a new idea and method for monitoring the healthy growth of vegetation.