一种基于TV-Gabor模型的高分辨率遥感图像农田信息提取方法  

Cropland Information Extraction from High Resolution Remote Sensing Image Based on TV-Gabor Model

作　　者：黄秋燕[1,2,3] 肖鹏峰[1] 冯学智[1] 吴桂平[4] 

机构地区：[1]南京大学地理信息科学系,南京210023 [2]广西师范学院北部湾环境演变与资源利用省部共建重点实验室,南宁530001 [3]广西师范学院资源与环境科学学院,南宁530001 [4]中国科学院南京地理与湖泊研究所,南京210008

出　　处：《遥感信息》2014年第2期77-82,88,共7页Remote Sensing Information

基　　金：国家高技术研究发展计划(2008AA12Z106);国家自然科学基金(40801166)

摘　　要：基于像元光谱特征的图像分类方法忽视对地物空间结构信息的利用,分类结果出现椒盐现象,难以满足高精度的农田信息提取要求。受人类视觉对图像几何结构、图像纹理边缘高敏感响应的感知机制启发,提出了一种基于TV-Gabor模型的高分辨率遥感图像农田信息提取新方法。首先分析遥感图像频谱特征获取图像先验信息,再结合TV-Gabor模型,提取遥感图像的几何结构特征,并将其用于改善基于像元光谱特征的农田信息提取精度。利用南京地区的QuickBird数据,验证了TV-Gabor模型提取几何结构特征对于提高高分辨率遥感图像农田信息提取精度的有效性。Image classification method based on pixels,which ignores the spatial feature of objects from high resolution remote sensing image,has salt and pepper phenomenon in its classification result,so it is difficult to meet the requirement of cropland information extraction.Inspired by the perception mechanism of human vision on image geometric structure and image texture,a new method of cropland information extraction from high resolution remote sensing image is proposed,which is based on TV-Gabor model with the image prior knowledge.First,geometry characteristics are extracted from high resolution remote sensing image based on the proposed model.Then,these geometry characteristics are applied to improve the precision of cropland information extraction by classification method based on pixels.Using QuickBird data in Nanjing,experimental results shows that geometry characteristics extracted by TV-Gabor model can effectively improve the accuracy of field information extraction from high resolution remote sensing image.

关 键 词：TV-Gabor模型 高分辨率遥感图像 几何结构特征 图像分解 农田信息提取 

分 类 号：TP701[自动化与计算机技术—检测技术与自动化装置]