中文摘要：

以SAR与SPOT多光谱影像为数据源，通过几何校正、辐射定标、降噪等数据预处理，选取PCA（Principal Component Analysis）、BT（Brovey Transform）、WT（Wavelet Transform）和HPF（High—Pass Filtering）4种方法进行融合实验，并对融合后数据以及原SPOT多光谱数据进行道路提取研究。结果表明：基于PCA和HPF方法所提取的道路特征明显，纹理清晰，错分和漏分现象不明显，道路信息效果较好；WT方法提取效果次之；BT道路提取效果相对较差；4种方法经融合后数据提取道路精度均高于原始SPOT多光谱数据，最优方法提取道路精度错分率较SPOT多光谱高达29．4％，道路提取精度相对较差的BT方法较SPOT多光谱错分率高出18％。

英文摘要：

With the two different remote sensing data sources （ SAR and SPOT multi-active spectrum ）, and through data preprocessing such as geometry correction, radiation correction and noise reduction, the paper makes a comparative analysis of the data fusion technology based on PCA （ Principal Component Analysis ）, BT （ Brovey Transform ）, WT （ Wavelet Transform ） and HPF （ High-Pass Filtering ）, then extracts road from fused image and SPOT multi-active spectrum. The result shows that PCA and HPF have the best effect, where the vein is clear, the rate of wrong classification and miss classification are lower, WT comes the second, and BT has poor effect relatively. The original SPOT image does not extract road effectively, its wrong classification is 29.4% higher than that of PCA, and 18% higher than that of BT.