论文部分内容阅读
利用GS变换、主成分分析、Ehlers变换、Wavelet分析、HIS变换5种方法对城区World View-2和PL-1A影像进行融合,并从影像融合质量和绿地信息提取精度两方面对融合方法的有效性进行了评价。结果表明:(1)5种融合方法中,GS变换融合的效果最好;主成分分析和Ehlers变换融合World View-2质量较好,但融合PL-1A影像质量较差;Wavelet变换、HIS变换融合两种影像质量都较差;(2)用于绿地信息提取时,GS、PCA融合影像获取的精度最高,其次为Ehlers、Wavelet融合影像,均明显高于多光谱影像的提取精度;Ehlers、Wavelet变换精度最低,绿地信息提取精度低于多光谱影像的提取精度。可以得出,影像融合可以明显地提高绿地信息提取精度,5种影像融合方法中,GS变换普适性较好,影像融合质量最好,提高分类精度效果最明显。
The fusion of World View-2 and PL-1A images in urban area was integrated by five methods: GS transform, principal component analysis, Ehlers transform, Wavelet analysis and HIS transform. The fusion method was validated from both image fusion quality and green space information extraction accuracy Sex was evaluated. The results show that: (1) GS fusion is the best among the five fusion methods; the quality of World View-2 fusion is better with principal component analysis and Ehlers transform, but the quality of fused PL-1A image is poor; Wavelet transform, HIS transform (2) For green space information extraction, GS and PCA fusion images have the highest accuracy, followed by Ehlers and Wavelet fusion images, which are significantly higher than the multi-spectral image extraction accuracy; Ehlers, Wavelet transform has the lowest accuracy and the accuracy of green space information extraction is lower than that of multi-spectral image. It can be concluded that the image fusion can obviously improve the accuracy of green space information extraction. Among the five image fusion methods, the GS transform has good universality, the best image fusion quality and the most obvious improvement of classification accuracy.