为提高CCD错位成像系统成像质量,提出用MTF方法定量评估CCD错位成像的成像质量,推导了CCD错位成像两种模式的MTF,取人眼能分辨的最低对比度0.05为阈值,分析表明,交错采样和四点采样模式的理论极限分辨率分别提高到1.4倍和1.86倍,奈奎斯特频率处MTF值分别提高了0.1106(27%)和0.1679(41%).用空间频率范围(0,0.5)内MTFA评估CCD成像质量,结果表明,交错采样模式和四点采样模式成像质量均优于CCD普通模式,且四点采样模式比交错采样模式成像质量进一步提高.建立了Matlab仿真CCD错位成像的数学模型,鉴别率板仿真结果验证了应用MTF定量评估CCD错位成像系统成像质量的正确性.

To improve image quality of subpixel imaging further, quantitative evaluation of image quality of subpixel imaging system based on MTF was proposed. The MTF of two main modes of CCD subpixel imaging was derived. Taking minimum contrast ratio of human eye 0.05 as threshold, analysis showed that theoretical limiting resolution of interlaced sampling mode and four-point sampling mode were improved to 1.4 times and 1.86 times respectively, and MTF values at Nyquist frequency of two modes were increased by 0.110 6 (27%) and 0.167 9 (41%), respectively. Evaluation of MTFA in (0, 0.5) illustrates that two CCD subpixel imaging modes both have better image quality than CCD common imaging mode, and the image quality of four-point mode is the best. A mathematical model for simulating CCD subpixel imaging using Matlab was established, the results of simulation of bar-target verify the correctness of the analysis based on MTF.

针对合成孔径雷达实时成像处理中数据量大、数据吞吐率高、成像算法实现复杂等特点，设计了适合于无人机载 SAR 实时信号处理系统的硬件平台和实时信号处理算法流程。该信号处理系统包括一块带有 AD 采集功能的接口板和两块以 TS201为核心处理器的信号处理板。考虑到实时性要求和无人机平台的不稳定性，设计了一种结合惯导和回波数据进行运动补偿的改进型 RD 成像算法。在无人机平台上成功稳定地实现大面积连续实时成像，证明信号处理系统稳定可靠，实时信号处理算法可行。

For solving the SAR imaging problems in real-time processing,such as huge data volume, high IO bandwidth and high algorithm complexity,a real-time signal processing flow and a hardware structure for UAV-borne SAR is designed.The hardware structure consists of an interface board with function of AD sam-pling and two signal processing boards based on TS201.Considering the real-time demand and the instability of UAV platform,an advanced RD algorithm is proposed in which motion error is compensated by both GPS and signal echo.The large area of consequent real-time SAR images are given on the UAV platform,which demonstrates the reliability of the processor and the feasibility of the real-time SAR imaging algorithm.

声子晶体的负折射成像已成为当今国内外研究的热点问题.首先通过分析弹性波激励声子晶体元胞的局域共振响应特性,计算了局域共振型声子晶体的带隙特性,以此为基础构建了声子晶体相应的实验模型.然后基于Comsol仿真平台,把声子晶体元胞在XY平面内排列成声子晶体板后,模拟了近场点声源通过声子晶体板的负折射声成像过程.同时计算了模拟两个理想点声源经过声子晶体板后负折射成像的特性.仿真结果表明:声子晶体能较好地分辨两个点声源所成的像,点像的最小分辨率可达到0.2λ,很好地突破了声学成像中的衍射极限.

The negative refraction imaging of phononic crystal has become a research focus at home and abroad .Localized resonant resonance characteristics of elastic wave excitation phononic crystal cell were analyzed ,and the phonon crystal band gap was calculated ,on this basis ,corresponding ex-perimental model of phononic crystal was built .Based on comsol simulation platform ,the phononic crystal cell were arranged as phononic crystal plate in XY plane ,and negative refraction acoustic im-aging process was simulated when near field acoustic source past through phononic crystal plate .Nega-tive refraction imaging characteristics was obtained when two ideal point sound source past through the phononic crystal plate .Simulation results show that phononic crystals can distinguish better the image which is made by two sound source ,the minimum resolution can reach 0 .2λ,so it can make a breakthrough in the diffraction limit of the acoustic imaging .

在工业CT成像应用中,板状物结构件受形状、结构、组分等特征的限制,使得射线能量与有效厚度不匹配,常规的扫描方式易使投影数据缺失,投影数据完备性较差。为此,讨论了一种基于倾斜圆扫描轨迹和代数迭代重建技术的CT成像模式。首先,在倾斜圆轨迹的基础上进行投影几何关系建模,结合空间几何知识完成投影矩阵的刻画;其次,基于迭代重建算法对轨迹的无约束化,研究任意扫描轨迹的迭代重建算法;最后通过板状物成像仿真实验,验证了算法的正确性和可行性。对于板状物结构件,该方法相对于传统的圆轨迹扫描,投影数据更完备,CT重建质量高。

In the application of industrial CT imaging, the X-ray energy mismatch to the effective thickness and traditional scanning is difficult to preserve the data integrity of projection due to the limit of tabular-struc-ture''s shape, structure and so on.In order to make up for the deficiency of projection data, a method based on the cone beam inclined circle scanning and the algebraic reconstruction technique ( ART ) is proposed.First of all, in the base of inclined circle scanning, the knowledge of analytic geometry with the projection is combined to obtain the projection matrix.The second, the iterative reconstruction algorithm is not binding to some trajec-tory, so we study the iterative reconstruction algorithm along a general scanning trajectory.Finally, according to the numerical simulation of tabular-structure the correctness and feasibility of the algorithm is verified.For the tabular-structure, compared to the traditional circle scanning, this method is better for preser

提出了一种利用相位拼接技术以扩大数字全息成像视场的方法。通过平移被测物体,记录物体各子区域的离轴全息图,然后利用菲涅尔再现算法得到部分物体相位像。在此基础上,采用均化误差算法对各子区域的相位像进行拼接,并对拼接相位像进行去倾斜处理,得到无畸变的扩展物体相位像。最后,利用USAF 1951反射型分辨率板对该方法进行实验验证。结果表明该方法能够克服CCD阵列大小对成像视场的限制,在保证成像分辨率的同时,有效扩展了数字全息的成像或测量范围。

A method to enlarge the field-of-view in digital holography imaging based on sub-aperture phase stitching tech-nique is proposed in this paper. The part information of an object is recorded by the off-axis digital holographic interferometer and the part phase image of the object is picked up by means of the Fresnel diffraction algorithm. Then the phase profiles in all the sub-regions are stitched together with the error averaging method,while the tilt of the stitched map are calculated and elimi-nated to obtain the orthoscopic profile of the whole surface. It was validated by the reflective resolution plate USAF 1951. The proposed method breaks away from the limits of the area of the CCD and expands the digital holography imaging field-of-view while the resolution is kept in the same level as that of the traditional methods.