植物叶片SPAD值反映了植物叶绿素含量, 对特定的植物也反映了氮含量。 为了实现植物叶绿素含量的快速无损检测, 利用光纤反射光谱技术对植物叶片SPAD值进行了预测建模研究。 实验中选取70个样本作为建模集, 50个样本作为校验集。 通过叶片光谱比对, 发现光谱红边段650～750 nm对SPAD预测建模有直接关系。 实验确定了光强调节因子和叶片厚度影响因子。 首先通过待定系数法构造出SPAD预测公式, 然后用Visual Basic6.0设计的遗传算法进行参数寻优, 最后确定最佳敏感波段为683.24～733.91 nm。 分析表明, 叶片厚度对SPAD反射光谱模型精度有显著影响。 经过叶片厚度修正后的建模集拟合因子Ｒ２为0.865 8, 校验集拟合因子Ｒ２为0.916 1。 结果表明, 利用反射光谱技术建立的SPAD预测模型是成功的, 从而可为仪器开发提供方法指导。

Handheld SPAD meter is often used to measure chlorophyll content of plant and nitrogen level for some species. For plant production automation, however, it loses its popularity due to its point-by-point checking. The authors need to monitor the growing conditions of plant remotely, instantly and nondestructively. In the test, we examined optical fiber reflection spectroscopy used to measure chlorophyll content of some plant leaves, or for their SPAD prediction. The authors picked 120 leaves randomly from our campus ground or trees, among which 70 samples were chosen as calibration set and others as verification set. Each sample was water-cleaned and air-dried. To locate each measuring point precisely when using SPAD meter and spectrometer, the authors drew a circle with a diameter of 10 mm on each leave to be measured. By comparing the spectral curves of various leaves, the authors found that the spectral band between 650-750 nm was significant for SPAD modeling since this range of spectral data of leaves with the same SPAD reading was close to each other. It was showed that leave color was an unnecessary factor for SPAD prediction by reflection spectroscopy. Besides, the authors discovered that LED’s narrow spectral range used by SPAD meter should be concerned because optical fiber spectrometer has much more wide spectral range. Based on this awareness, the authors designed an adjustment factor of light to linearly rebuild spectrometer’s reflective intensity so that it reached zero outside the band 650-750 nm. Moreover, leave thickness was another influential factor for SPAD prediction since the light of SPAD meter goes through the leave while the reflective spectrometer does not. First, an equation for SPAD prediction was built with uncertain parameters. Then, a standard genetic algorithm was designed with Visual Basic 6.0 for parameter optimization. As a result, the optimal reflection band was narrowed within 683.24-733.91 nm. The result showed that leave thickness strongly affects the precision of SPAD prediction. Through the modification of leave thickness, the regression coefficient (R2) of calibration set and verification set reached 0.865 8 and 0.916 1 respectively. The test showed that optical fiber reflection spectroscopy is useful for SPAD prediction and can be used to develop remote SPAD sensor.