摘要
在一些遥感图像中由于目标不清晰或背景复杂,使得目标的检测变得困难,图像融合可以把来自多传感器的图像信息综合起来,提高对图像信息分析和提取的能力。通过把同一目标的不同传感器获得的图像数据利用小波包变换进行融合,这种方法能够很好地将多源图像的细节融合在一起,得到目标较为清晰的融合图像,在此基础上利用数学形态学的方法进行目标检测,得到了满意的效果,证明了这种方法的有效性。
It is difficult to detect targets in some remote sensing images because of blurred objects or complicated background. Image fusion can merge information from multi-sensor and improve the performance of information analysis and feature extraction. Through merging image data from multi-sensor to the same object by using wavelet packet analysis, a fused picture in which the target is clearer than that in the original images can be obtained. The details of input image can be fused successfully, and information of each input image can be displayed perfectly. Then mathematical morphology method is applied to detect the target based on fused picture and satisfactory results are gotten.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2001年第4期454-457,共4页
Pattern Recognition and Artificial Intelligence
基金
国家自然科学基金
国防科技预研基金
