| 摘要: |
| 穿墙雷达成像中，建筑物布局成像对墙内隐藏目标探测和提取有重要辅助意义，完整的建筑物布局成像大多在多视角下进行，对此提出一种基于多方位多尺度的多视角建筑物布局成像融合方法。首先由后向投影算法获得各视角墙体成像后，采用多方位二维匹配滤波对各视角图像增强,消除旁瓣、栅瓣的影响。然后结合多尺度变换，即非下采样Contourlet变换分解各个视角的图像，将获得的低频分量和高频分量分别采用相应的融合规则进行融合。最后通过非下采样Contourlet反变换获得最终融合后的建筑物布局图像。仿真实验和实测数据结果证明了该方法的有效性。 |
| 关键词: 穿墙雷达 多方位多尺度 建筑物布局成像 图像融合 |
| DOI：DOI: 10.3969/j.issn.1672-2337.2019.01.015 |
| 分类号:TN957.52 |
| 基金项目:国家自然科学基金（No.61461012）； 广西自然科学基金（No.2017GXNSFAA198050）； 广西无线宽带通信与信号处理重点实验室2016主任基金项目（No.GXKL06160106） |
|
| Multi-Azimuth and Multi-Scale Building Layout Imaging Fusion Method |
|
TANG Qian, JIN Liangnian
|
|
1.Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing, Guilin 541004, China;2.2.School of Information and Communication, Guilin University of Electronic Technology, Guilin 541004, China
|
| Abstract: |
| For through-wall-radar imaging，building layout imaging has important meaning for hidden target detection in the wall. The complete building layout imaging is carried out almost in multi-view. In this paper, a multi-azimuth and multi-scale image fusion method from multiple perspectives is proposed. First of all, the back-projection algorithm is used to obtain image of each view and the multi-directional two-dimensional matched filter is used to strengthen the single-view image and eliminate the effects of the sidelobes and grating lobes. And then, combined with multi-scale transform, the down-sampled Contourlet transform decomposes the images of each view, and the obtained low-frequency components and high-frequency components are fused respectively with the corresponding fusion rules. Finally, the final fused building layout image can be obtained by nonsubsampled Contourlet inverse transform. Simulation results and measured data prove the effectiveness of this method. |
| Key words: through-wall-radar multi-azimuth and multi-scale building layout imaging image fusion |
