基于高光谱成像技术的水稻叶片SPAD值及其分布问题研究

Research on SPAD value prediction and distribution of the rice leaf based on Hyperspectral imaging technique

高光谱成像技术在快速无损检测植物叶片叶绿素含量上得到越来越广泛的应用.运用SPAD仪可同期获得叶片的叶绿素含量.以水稻叶片为研究对象,首先采集水稻活体植株至培养皿,利用SPAD502叶绿素计采集叶片的SPAD值,最后使用高光谱成像仪采集水稻叶片的高光谱影像.运用不同的植被指数和偏最小二乘法分别对SPAD值进行回归分析.结果显示,偏最小二乘回归模型精度较高且较为稳定.根据最佳预测模型反演叶片上任意像素的SPAD含量,通过伪彩色配色即可得到水稻叶片SPAD分布图像.该方法为研究水稻植株的生长状况提供了更为具体的数据资料,为水稻的产量估测和病害预警提供了新的依据和方法.

Hyperspectral imaging technology gets more and more widely using in the fast and non-destructive testing of the plant chlorophyll content which can be obtained by SPAD.Firstly,living rice plants were collected to the dishes.Secondly,SPAD value of the rice leaves were measured by SPAD502 chlorophyll meter.Finally hyper-spectral images of rice leaves were captured from hyper-spectral sensor.Using different vegetation indexes and partial least squares regressed with SPAD value respectively.The results showed that accuracy of partial least squares regression model was higher and more stable.The SPAD value of each pixel in image was estimated by the fitted model,and the distribution of SPAD value in the rice leaves were described by pseudo-color map.The method provides more specific data,new basis and methods for studying the growth conditions of rice planting.