【作者】 王晓蕊；

【导师】 张建奇； 冯卓祥；

【作者基本信息】 西安电子科技大学 ， 光学工程， 2005， 博士

【摘要】 红外焦平面阵列技术的发展日趋成熟,使得新型焦平面成像系统以其独特的优点逐渐替代传统的扫描型成像系统,并将成为未来成像系统发展的主流趋势。新型红外焦平面成像系统由于引入了新工艺、新结构和新技术,使得影响系统性能的因素发生了明显的变化。而传统成像系统理论模型及性能表征方法对这些新技术考虑不足或没有考虑,因此,非常有必要改进传统的性能评价方法或选择全新的性能评价方法以适应不断发展的新型红外成像系统的性能评估。 针对新型红外焦平面成像系统,本论文开展了如下研究工作:(1)深入分析了焦平面成像系统引入的新技术及其导致的各种典型物理效应,以连续/离散/连续分析模型为基础,重点研究了非线性因素:光学扭曲、渐晕、采样效应、非线性响应特性、噪声、重构显示算法等的建模仿真方法,建立了较为完善的焦平面红外成像系统仿真模型,并利用最小可分辨率温差MRTD的实验仿真预测结果验证了仿真模型的合理性。(2)根据微扫描成像机理,推导了微扫描探测器阵列的象素传递函数模型,并利用此模型首次定量分析了探测器占空比、微扫描模式对微扫描成像质量和分辨率的改善程度,确定了特定占空比探测器的最优微扫描模式的选择方法;并利用最小可分辨温差模型和压缩传递函数,定量分析了微扫描模式对成像系统现场性能的改善程度,可为微扫描成像系统的优化设计提供理论支持。(3)立足于红外成像系统性能评价的新概念—三角形方向鉴别阈值(TOD)度量方法,设计制作了等边三角形方位靶标,利用四选一测量过程研究了三角形方位鉴别概率的测量方法,利用Weibull心理测量函数拟合鉴别概率数据,实现了TOD曲线的度量;并针对TOD试验测量平台和测量方法,首次对TOD度量结果的不确定度进行了分析,验证了TOD测量结果的稳定性。(4)在焦平面成像系统各模块建模的基础上,仿真生成三角形靶标图像,构建了三角形靶标方位分类器,建立了基于三角形靶标成像仿真的TOD预测理论模型。(5)以可感知信噪比为理论基础,利用匹配滤波器的概念首次建立了扫描型和凝视型红外成像系统的TOD理论模型,以试验数据为基础,验证了TOD性能理论模型的正确性;并利用TOD理论模型定量分析了系统结构参量对系统综合性能的影响,探讨了基于TOD理论模型的焦平面红外成像系统匹配设计方法 (6)最后,依据外场试验数据,确定了TOD曲线向外场性能转换的TOD准则;以TOD准则为理论基础,首次提出了一种TOD曲线与目标热结构分布特性有效结合的目标获取改进模型,能够更精确地预测系统的现场性能。更多还原

【Abstract】 As infrared focal plane array (FPA) detectors continue to mature, FPA based IR imaging systems become a main trend in infrared imaging field compared to scanning imaging system. How to effectively characterize the performance of infrared imaging system based on FPA become more and more important. Traditional performance characterization methods or models become insufficient for this type of imaging systems. Hence, it is necessary to improve traditional performance evaluation method or propose completely new performance model to satisfy the performance evaluation of the novel IR FPA imaging system. In this paper, the following works are studied.(1) According to signal transfer mechanism of IR FPA imaging system, the Continuous/Discrete/Continuous model is introduced to analyze various effects of infrared imaging system. Based on the modeling of the linear effect, nonlinear effects: geometric distortion, vignetting, sampling artifacts, nonlinear response character, noise, reconstruction algorithm, et al are modeled and simulated utilizing the pixel processing method in spatial domain. Further, the model of imaging system is validated by minimum resolvable temperature difference(MRTD) experiment simulation.(2) Pixel transfer function(PTF) used to describe micro-scanning imaging process is derived for the staring detector array. Further, this model is utilized to calculate the amount of the improvement of micro-scanning to the imaging resolution for the detector array of different fill factors. Analytical results show that micro-scanning can significantly improve the resolution and quality of the reconstructed image, and the amount of improvement of different micro-scanning modes to the imaging resolution is different, which is closely associated with the fill factor. It will be helpful to choose the optimum micros-canning mode according to the fill factor of the detector array. On the other hand, an application-oriented analytical method of micro-scanning imager is proposed. The modified model of MRTD is achieved by considering the micro-scanning. The field performance of infrared imaging system based on micro-scanning is predicted and discussed with the target acquisition model. The simulated results show whether micro-scanning improve infrared imaging system performance depends on the other imaging system parameters, and working conditions.(3) Based on the idea of Triangle Orientation Discrimination Threshold(TOD) measurement method, several triangle patterns with different sizes are designed, accurate TOD measure is studied. Uncertainty of the TOD measure curve is firstanalyzed quantitatively by considering the components of the experimental instrument and the measure method, and the stability of the TOD measure results is validated.(4) A method for predicting TOD curve is proposed based on the triangle pattern simulation image. Specifically, the standard equilateral triangle patterns with different spatial frequency and different radiometric temperature are first generated, and discriminated by triangle orientation classification algorithm. The correct discrimination probability of the triangle pattern orientation is obtained utilizing the statistical method. The psychometric function-Weibull function is used to fit the probability data, and the TOD prediction curve is achieved. Comparisons with the experimental results indicate that the TOD prediction curve is reasonable.(5) Equation-based TOD performance model was first developed focusing on staring and scanning thermal imagers. Specifically, the spatial distribution of the non-periodic triangle standard pattern is determined quantitatively. Various effects of overall components of system on the triangle standard pattern are analyzed and modeled. The matched filter idea is adopted to quantitatively characterize the spatial-temporal integration of the human visual system to signal, aliasing and various noise components over the triangle pattern area, and the perceived signal-to-noise ratio of thermal imager is derived. Experimental results show that this model can give reasonable prediction of the TOD performance curve for the scanning and staring imaging system.(6) According to the idea of TOD measure method, TOD criteria is determined utilizing the Weibull psychological function based on the field experimental data. A novel target acquisition (TA) performance prediction method based on the triangle orientation discrimination (TOD) threshold curve is proposed by considering thermal structure of the target. Comparisons with the experimental results show that this improved TA approach indeed leads to more accurate prediction of the field performance over the TA model based on the MRTD curve.更多还原