合成孔径声纳高效成像算法研究及其实时实现
文献类型：学位论文
| 作者 | 陈东升 |
| 学位类别 | 博士 |
| 答辩日期 | 2007-05-31 |
| 授予单位 | 中国科学院声学研究所 |
| 授予地点 | 声学研究所 |
| 关键词 | 合成孔径声纳 运动补偿 实时处理 并行成像 信号处理系统 |
| 其他题名 | Research on Efficient Image Reconstruction Algorithms and Real-Time System Realization for Synthetic Aperture Sonar |
| 学位专业 | 信号与信息处理 |
| 中文摘要 | 本论文对合成孔径声纳（Synthetic Aperture Sonar, 缩写为SAS）的多子阵快速成像算法、运动补偿以及基于多DSP并行实时SAS信号处理系统设计等三个方面的问题展开了深入的研究。 在SAS多子阵快速成像算法方面的研究包括以下两项内容：（1）分析了基本逐点成像算法、快速逐点算法及 算法的优化实现问题，通过仿真实验验证了算法的正确性并比较了几种算法实现的效率和内存占用开销，为实时SAS信号处理软件系统的设计提供了理论指导。（2）分析了方位向速度扰动对SAS成像质量的影响，提出了一种基于鲁棒卡尔曼滤波的数据平滑化的处理方法，对传感器测量值的野值及扰动有很好的抑制作用。 拖体的姿态对SAS图像质量的影响很大，运动补偿（motion compensation）技术是普遍采用的消除或抑制运动误差的有效方法。本文在运动补偿方面的研究主要是在DPC方法的基础上提出了一种改进的运动补偿方法，扩展了DPC算法的适用性。 本文还系统地研究了基于多DSP的实时SAS信号处理软件系统的设计，在并行实时成像方面所做的研究内容有：（1）通过分析SAS信号处理的功能模块组成，设计了一种基于多DSP的多级流水线并行处理结构，使数据处理实时、高效实现。（2）本文设计的SAS实时信号处理系统成功应用于863计划课题项目“合成孔径声纳海试样机”中，处理了大量的水池、湖试和海试实验数据，给出了部分的SAS成像结果。 |
| 英文摘要 | In this dissertation three key aspects of Synthetic Aperture Sonar (SAS) are discussed in detail, in which the multi-sub-array fast imaging algorithms, motion compensation and parallel real-time signal processing systems based on multi-DSP are included. As far as the multi-sub-array fast imaging algorithm is concerned, two contents are studied. Firstly, the optimized realizations are investigated for fundamental pixel-based imaging algorithm, fast pixel-based imaging algorithm, and the ω-k algorithm. The efficiency and memory costs of these algorithms are compared through simulation experiments, and the validity of these algorithms is also verified by these experiments. All these works provide theoretical guidance for the design of real-time SAS signal processing software system. Secondly, the issue is analyzed about how the advance velocity disturbance influences the imaging qualities. Furthermore, a data smoothing approach is proposed based on robust-Kalman filtering, which satisfyingly get rid of the observation outlier and depresses disturbances to the sensors. In SAS research, the attitude of the towfish influences the imaging quality greatly, and motion compensation is a widely used effective approach to depress the motion errors. In this dissertation, a modified motion compensation method is proposed based on the displaced phase center (DPC) algorithm, which extends the applicability of DPC algorithm. The implementation procedures and steps of this modified algorithm are also presented in this dissertation, and the efficiency is also verified through simulations and the sea trial data. The design of parallel real-time SAS signal processing systems based on multi-DSP is also systemically studied in this dissertation. The parallel real-time imaging research includes: (1) through analyzing the structure of SAS signal processing modules, a multilevel pipelining parallel processing structure based on multi-DSP is designed, which ensures real-time and efficient signal processing. (2) The real-time SAS signal processing system designed in this dissertation has been applied in the ‘SAS prototype machine for sea trial’ which was funded by the ‘863’ project, and lots of experimental data from lake trials and sea trials has been processed using this system, and some of the SAS imaging results are presented. At the end of the dissertation, the whole research work is summarized. |
| 语种 | 中文 |
| 公开日期 | 2011-05-07 |
| 页码 | 120 |
| 源URL | [http://159.226.59.140/handle/311008/162]  |
| 专题 | 声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文 |
| 推荐引用方式 GB/T 7714 | 陈东升. 合成孔径声纳高效成像算法研究及其实时实现[D]. 声学研究所. 中国科学院声学研究所. 2007. |
入库方式： OAI收割
来源：声学研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。