随着精准农业这一理念的提出, 快速精确获取农业信息成为人们关注的重点。 偏振遥感因其结合了多角度遥感的特点, 高光谱遥感和微波遥感的特点, 从而能在不破坏地物的基础上提高探测和识别地物的准确度。 以往的研究主要以单一土壤肥力指标为主。 该实验通过实测吉林省典型地区农田土壤在不同状态下的光谱曲线, 探讨了在确定最佳观测条件下的土壤肥力综合指标IFI与土壤光谱曲线的关系。 结果表明光线探测天顶角、 探测方位角、 偏振状态均会在一定条件下影响土壤光谱曲线, 但对光谱的形状和走向影响不大。 在保证设备可达条件下, 设计出遥感中监测土壤肥力的最佳方式。 通过对光谱数据进行一阶微分和倒数的对数运算, 在选定的特征波段上建立相应的肥力模型。 结果表明土壤肥力与光谱反射比呈现明显的负相关关系, 和光谱吸光度呈现明显的正相关关系, 与光谱反射比的一阶微分相关关系不确定。 在利用光谱反射比、 光谱一阶反射比和吸光度特征参数进行土壤肥力估算时, 发现光谱反射比与土壤肥力的二次函数拟合优度最佳, 在560和860 nm处决定系数分别达到0.876和0.867。

With the introduction of the concept precision agriculture, acquiring information quickly and precisely has become the focus researchers. Polarization remote sensing can improve the accuracy of exploring and identifying features without causing damage, because it colligates multi-angle remote sensing, hyper-spectral remote sensing and microwave remote sensing. Previous researches are mainly based on single standard of soil fertility. This study explores the relationship between Integrated Fertility Index (IFI) and soil spectrum curve under the best observation conditions through measuring the spectrum curves in different conditions of typical farmland soil in Jilin Province. The study shows that incidence angles, relative azimuth angles and polarized states will affect soil spectrum curve under certain circumstances In fact, the design of the most workable remote sensing to observe soil fertility for this purposecan be realized by equipment. In this way, soil reflection was mathematically manipulated into first derivative reflectance spectra and inverse-log spectra, then established soil fertility model on ground of characteristic bands. The study shows the negative correlation between soil fertility and reflectance, positive correlation between soil fertility and absorptivity of spectrum, but the correlation between soil fertility and reflectance of spectrum in first differential is uncertain. When using reflectance of spectrum and first differential reflectance of spectrum and absorbance indexes to calculate soil fertility, it has been found that the quadratic function of reflectance of spectrum and fertility fits properly , with determination coefficient reaching 0.876 and 0.867 at 560 and 860 nm respectively.