建筑物透视探测场景成像算法研究
发布时间：2019-07-08 15:53
【摘要】：出于维稳、侦查和救援等需要,穿墙雷达技术逐渐成为研究热点。建筑物透视探测场景成像技术以穿墙雷达成像技术为依托,主要实现建筑布局成像和隐蔽目标成像两个功能,分别用来形成建筑布局全景图像和隐蔽目标图像。穿墙雷达合成孔径成像技术按探测模式可以分为条带SAR模式和聚束SAR模式。传统条带SAR模式多采用多视角探测,但实际应用的环境复杂性通常不具备多视角探测的条件。另外,电磁信号穿墙传播引起的信号强度衰减带来建筑布局成像强度的非均匀性,这为最终建筑布局的检测和获得带来了较大的困难。再者,在建筑物的封闭探测环境中,墙体造成的目标多径回波干扰严重,成像过程中产生大量的多径幻象,无法形成高质量的隐蔽目标图像。本文针对以上这些问题,做了以下工作:1、针对车载聚束SAR模式能在单一视角一次性完成场景成像,适合复杂环境下探测的特点,研究了车载聚束SAR模式的信号模型,给出该模式下的RMA和BP成像算法;2.提出一种基于模糊逻辑的PCA图像增强方法,有效抑制杂波和噪声并增强了建筑布局图像强度,实测数据处理验证了算法的有效性;3.提出一种基于霍夫变换门限检测的多层墙体成像衰减补偿算法,有效补偿了单视角探测下建筑布局图像强度的差异性;4.针对不同视角下的多幅建筑布局图像,提出门限平均加权融合、自适应加权融合、像素选大融合和小波变换融合等图像融合方法,形成一幅更清晰、信噪比更高的全景图像;5.提出一种基于子孔径成像的多径幻象抑制方法,在单视角探测条件下利用目标多径的“方向依赖”特性有效抑制了多径幻象目标,提升了隐蔽目标图像信噪比;6.在多视角探测条件下,针对隐蔽目标多径幻象干扰问题,分析幻象成因机理,将不同视角下的图像作相乘融合,大大抑制了多径幻象,提升了隐蔽目标图像信噪比。
文内图片：
图片说明：左图为原始回波数据，，右图为距离像平面
[Abstract]:Through the need of stability, investigation and rescue, through-wall radar technology has become the focus of research. The building perspective detection scene imaging technology is based on the through-wall radar imaging technology, which mainly realizes the two functions of the building layout imaging and the concealed target imaging, and is used for forming the building layout panoramic image and the hidden target image respectively. The detection mode of the through-wall radar synthetic aperture imaging technology can be divided into the strip SAR mode and the spotlight SAR mode. The traditional stripe SAR mode is multi-angle detection, but the environment complexity of the practical application is usually not the condition of multi-view detection. In addition, the signal intensity attenuation caused by the through-wall propagation of the electromagnetic signal brings about the non-uniformity of the imaging intensity of the building layout, which brings great difficulty to the detection and the acquisition of the final building layout. Furthermore, in the closed detection environment of the building, the target multipath echo interference caused by the wall body is serious, a large number of multi-path phantom is generated in the imaging process, and the high-quality hidden target image cannot be formed. In view of the above problems, the following work is done:1. For the vehicle-mounted spotlight SAR mode, the scene imaging can be done at one time in a single visual angle, which is suitable for the detection in a complex environment. The signal model of the vehicle-mounted spotlight SAR mode is studied, and the RMA and BP imaging algorithm in this mode are given. 2. A method of PCA image enhancement based on fuzzy logic is proposed, which can effectively restrain the clutter and noise and enhance the image strength of the building layout. A multi-layer wall imaging attenuation compensation algorithm based on the Hough transform threshold detection is proposed, which can effectively compensate the difference of the image intensity of the building layout under the single-view angle detection. Aiming at the images of multiple buildings in different visual angles, an image fusion method such as a threshold average weight fusion, a self-adaptive weighted fusion, a pixel selection fusion and a wavelet transform fusion is proposed to form a panoramic image with a clearer and higher signal-to-noise ratio; and 5. A multi-path phantom suppression method based on sub-aperture imaging is proposed, and the multi-path phantom object is effectively suppressed by the "direction dependence" characteristics of the target multipath under the single-view angle detection condition, and the signal-to-noise ratio of the hidden target image is improved; Under the condition of multi-view detection, aiming at the problem of multi-path phantom interference of the hidden target, the mechanism of the formation of the phantom is analyzed, the images under different views are combined and fused, the multi-path phantom is greatly reduced, and the signal-to-noise ratio of the hidden target image is improved.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
本文编号：2511702
文内图片：
图片说明：左图为原始回波数据，，右图为距离像平面
[Abstract]:Through the need of stability, investigation and rescue, through-wall radar technology has become the focus of research. The building perspective detection scene imaging technology is based on the through-wall radar imaging technology, which mainly realizes the two functions of the building layout imaging and the concealed target imaging, and is used for forming the building layout panoramic image and the hidden target image respectively. The detection mode of the through-wall radar synthetic aperture imaging technology can be divided into the strip SAR mode and the spotlight SAR mode. The traditional stripe SAR mode is multi-angle detection, but the environment complexity of the practical application is usually not the condition of multi-view detection. In addition, the signal intensity attenuation caused by the through-wall propagation of the electromagnetic signal brings about the non-uniformity of the imaging intensity of the building layout, which brings great difficulty to the detection and the acquisition of the final building layout. Furthermore, in the closed detection environment of the building, the target multipath echo interference caused by the wall body is serious, a large number of multi-path phantom is generated in the imaging process, and the high-quality hidden target image cannot be formed. In view of the above problems, the following work is done:1. For the vehicle-mounted spotlight SAR mode, the scene imaging can be done at one time in a single visual angle, which is suitable for the detection in a complex environment. The signal model of the vehicle-mounted spotlight SAR mode is studied, and the RMA and BP imaging algorithm in this mode are given. 2. A method of PCA image enhancement based on fuzzy logic is proposed, which can effectively restrain the clutter and noise and enhance the image strength of the building layout. A multi-layer wall imaging attenuation compensation algorithm based on the Hough transform threshold detection is proposed, which can effectively compensate the difference of the image intensity of the building layout under the single-view angle detection. Aiming at the images of multiple buildings in different visual angles, an image fusion method such as a threshold average weight fusion, a self-adaptive weighted fusion, a pixel selection fusion and a wavelet transform fusion is proposed to form a panoramic image with a clearer and higher signal-to-noise ratio; and 5. A multi-path phantom suppression method based on sub-aperture imaging is proposed, and the multi-path phantom object is effectively suppressed by the "direction dependence" characteristics of the target multipath under the single-view angle detection condition, and the signal-to-noise ratio of the hidden target image is improved; Under the condition of multi-view detection, aiming at the problem of multi-path phantom interference of the hidden target, the mechanism of the formation of the phantom is analyzed, the images under different views are combined and fused, the multi-path phantom is greatly reduced, and the signal-to-noise ratio of the hidden target image is improved.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
本文编号：2511702
本文链接：https://www.wllwen.com/kejilunwen/wltx/2511702.html
