从地物光谱图像识别的关键问题出发, 将改进的梯度方向直方图(光谱梯度方向直方图)与“光谱总反射率”应用到偏振光谱识别研究中, 实现了对5种干枯植物与3种裸土的识别.在研究中, 先联合光谱梯度方向直方图特征向量的模与光谱总反射率进行层次聚类分析, 可以识别6种研究对象, 再单独利用光谱梯度方向直方图特征的模进行层次聚类分析识别其余2种研究对象, 从而可以实现全部研究对象偏振光谱识别.

Starting from the key problems of spectral image recognition, the improved Histogram of Oriented Gradient (Spectral Histogram of Oriented Gradient) and "Spectral total reflectance" were applied to the polarized spectral recognition study, and five dry plants and three bare soils could be recognized. In the study, the hierarchical clustering analysis was carried out by combining the length of Spectral Histogram of Oriented Gradient and Spectral total reflectance, and six study objects could be identified. Then, the length of Spectral Histogram of Oriented Gradient was used to identify the remaining two study objects by the hierarchical clustering analysis alone. As a result, the polarized spectrum of all the study objects could be recognized.