摘要
为了更有效地实现目标识别 ,采用了像素级图像融合方法 ,该方法简单 ,易于实现 ,其中低通对比度金字塔融合和多尺度对比度增强融合采用金字塔法实现 ,实验结果显示多尺度对比度增强的融合效果优于前者 .但采用金字塔法有一个不可避免的缺陷 ,就是在做子采样时由于频率叠混效应而可能引进噪声 .而非金字塔式分层融合技术则避免了金字塔法的这种缺陷 。
Image fusion is preferrable to ATR. Pixel level fusion algorithms can be realized simply. One scheme is for contrast or ratio of low pass pyramid, and another one is for multi scale contrast enhancement, both being based on constructing pyramid. The experimental results indicate that the latter is more powerful than the former. But the algorithms based on pyramid scheme are affected with effect of frequency aliasing due to sub sample and the noise is added to fused image. So another algorithm, namely non pyramid hierarchical fusion scheme is presented in order to get rid of this defect.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2002年第5期25-27,共3页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目 (6 0 0 85 0 0 2 )
国防科技预研基金资助项目
