土壤的光谱特征及氮含量的预测研究

应用近红外光谱分析技术(NIR)测定土壤参数具有快速、方便的特点.文章分析了不同含水率、不同颗粒大小的土壤样本在不同测试角、不同测试高度对土壤光谱的影响,并得到了不同含水率和不同粒径土壤的含氮量预测模型.研究了这些因素对含氮率测量的影响,分析了NIR技术在田间实地应用预测的可能性.研究表明,光谱仪在距土壤高度为100 mm,测试角为45°时,具有最大的吸光度.土壤粒径和含水率这2个参数明显影响,当粒径在0.5～5 mm变化时,含氮量预测相关系数r为0.81左右,当土壤粒径在＜0.25和＞5 mm模型的预测能力变差.当土壤样品呈天然潮湿状态时,氮的预测结果较差.而样品干燥以后,预测相关系数较高.为土壤原位光谱测试提供了依据.

Spectral Characterization and N Content Prediction of Soil with Different Particle Size and Moisture Content

The present work was focused on analyzing the influence of moisture content, particle size, light source incidence angle and observation height on a loamy mixed soil spectra Meanwhile, prediction models for N content with different moisture and particle sizes were obtained, and the influence of these properties on N prediction was studied. The future applicability of NIR spectroscopy as a technique able to make prediction on the spot was analyzed. Observation height 100 mm and light source angle 45 degrees were chosen to present a sharpest spectra. Moisture content and particle size were found to affect strongly the absorbance of the spectra, and an accurate N prediction was obtained when the particle sizes varied from 0. 5-1. 0, 1. 0-2.0 and 2-5 mm with r of 0. 82, 0. 81 and 0. 81, respectively. Poor N prediction was obtained when the soil kept its natural moisture with r of 0. 57 and SECV of 3. 06 compared with the performance when it was dry with r of 0. 81 and SECV of 2. 40.