摘要
针对àtrous小波分解过程中的细节损失以及融合结果边缘模糊等问题,提出一种基于高提升滤波和àtrous小波分解的遥感图像融合算法.首先,利用àtrous小波变换得到第一层小波面,然后将获取的小波面作为非锐化模板进行高提升滤波,最后将滤波后的影像作为高频分量参与àtrous小波重构得到融合图像.基于北京地区的高分一号影像进行了算法验证,结果表明本算法优于传统小波域中的图像融合算法,对于城镇等建筑密度比较高的地区,融合后的影像保留了更多的细节信息,可以更好地支持城市变化检测和城市土地利用分析等方面的应用研究.
Remote sensing image fusion is widely used in object recognition because of complementary nature of images from different sensors. However, in the fusion process of the traditional àtrous wavelet transform, there are some problems, such as loss of details and blurred image edge of fusion image. The analysis of the wavelet planes showed that these problems are caused by the detail loss of the low-pass filtering in the àtrous wavelet transform. A novel remote sensing image fusion algorithm was proposed, based on high boost filtering and àtrous wavelet transform. First, the first wavelet planes were obtained by àtrous wavelet transform. Then, the first wavelet plane was treated as a sharpening template to participate in the high-boost filtering, so the image details can be enhanced. Finally, the high resolution multi-spectral image was obtained from the reconstruction of wavelet using the filtered image. The algorithm was tested on GF-1 image of Beijing. The results show that our algorithm can reserve more detail information of images with a high building density. This algorithm is helpful for the research such as the change detection of city or the analysis of the utilization of the urban land.
出处
《河南大学学报（自然科学版）》
CAS
2016年第2期202-206,共5页
Journal of Henan University:Natural Science
基金
国家高分重大专项课题(Y4D0100038)
国家973计划课题(Y070072070)
中科院战略性先导专项子课题(Y1Y02230XD)
