论文部分内容阅读
本文利用尺度空间理论对高光谱遥感数据中包含的精细光谱信息进行多尺度观察,在特定的尺度层次下提取能够对地物类别属性做出判断的定性约束特征,在此基础上结合光谱相似性测度最终确定像元所属类别。试验结果表明该方法可有效减少传统匹配算法由于噪声、成像环境等因素引起的误判、错分问题,提高分类识别的精度。
In this paper, we use the scale space theory to observe the detailed spectral information contained in hyperspectral remote sensing data at multiple scales and extract the qualitative constrained features that can be used to judge the attributes of the object categories at a particular scale. Based on this, we combine the spectral similarity The measure finally determines the category to which the cell belongs. The experimental results show that this method can effectively reduce the misjudgment and misclassification problems caused by the traditional matching algorithms due to noise, imaging environment and other factors, and improve the accuracy of classification and recognition.