基于知识发现和决策规则的遥感图像城区土地覆盖/利用分类方法

Remote Sensing Image Urban Land Cover /Use Classification Based on Knowledge Discovery and Decision Rule

提出一种利用地类的色度信息、小波提取纹理特征、植被指数及形状知识等采用规则推理有效识别遥感土地类别的方法。采用知识发现和决策规则方法,可以充分吸收遥感专家的思想和工作经验,可以充分利用多种模糊性地学知识来提高遥感影像分类效果。最后通过实例证明了该方法的有效性。

This paper presents the remote sensing image land cover /use classification method,based on rule-reasoning method using gray information,wavelet-based texture feature,vegetation index and shape information.The method based on knowledge discovery and decision rule,may sufficiently mine remote sensing expert thoughts and experiences and much fuzzy geo -knowledge to improve remote sensing image classification result.In the study,perfection knowledge discovery rules are constructed to discriminate land cover /use type in Xuzhou urban.