摘要
针对传统的影像分辨率增强算法存在的缺点,提出了一个顾及影像纹理特性的自适应分辨率增强算法。在算法中,首先通过影像分析,根据定义的3种纹理特性,确定出像素在影像中所在的位置是属于平滑纹理、一般纹理还是边缘纹理,然后根据不同的属性应用不同的滤波器进行内插增强。特别是对边缘纹理的处理,为了提高精度,是通过两次迭代来完成的。实验中利用该算法和传统的增强方法对三峡地区航空影像和武汉TM影像分别进行了处理,并用常用的影像质量评价方法对增强处理结果进行了对比评价,实验证明,所提出的算法在保持纹理、边缘方面有较大的优越性,具有实用价值。
Image resolution enhancement is useful in many applications.However,traditional image resolution en-hancement algorithms often produce images with various artifacts,especially around edges on the image.In this paper,contraposing the disadvantages of traditional algorithms,we present an adaptive algorithm to enhance image resolution.In this algorithm,the texture attribute of images is considered.Appropriate filters are selected in different texture circum-stances.Especially for the edge positions,the resolution enhancement is completed in two iterations in order to improve the precision and try to interpolating along edges,not across them.Therefore,this algorithm is based on edge detection technique.Through a MSE evaluation comparison,this algorithm is better than traditional interpolation methods(quadratic and cubic interpolation functions).This algorithm does not suffer from ringing and blurring artifacts.It is demonstrated that our algorithm is reasonable and effective.
出处
《遥感学报》
EI
CSCD
北大核心
2005年第3期253-259,共7页
NATIONAL REMOTE SENSING BULLETIN
基金
国家自然科学基金(40471088)
中国973发展计划(2003CB415205)
中国科学院遥感科学国家重点实验室开放基金
地理空间信息工程国家测绘局重点实验室开放基金。
关键词
分辨率
增强
内插
纹理
resolution
enhancement
interpolation
texture
