红外探测器非均匀校正算法研究

发布时间：2018-10-15 17:42

【摘要】：红外图像的非均匀性是伴随着红外焦平面阵列(Infrared Focal Plane Arrays,IFPA)一同出现的,非均匀性在红外图像中表现为固定图案噪声,严重限制了红外成像系统的成像质量。目前国际上关于红外图像非均匀校正的方法主要有基于黑体定标的方法和基于场景校正的方法。本文首先介绍了红外成像系统的组成和红外探测器的发展,并介绍了红外图像非均匀性校正方法的分类和发展。然后介绍了非均匀性的定义和来源,以及红外探测器的两种响应模型。同时,介绍了几种典型的非均匀性校正算法,分析了各种方法的基本原理和处理步骤以及优缺点。接着介绍了各种图像配准方法的基本原理和处理步骤以及优缺点,并详细介绍了基于图像配准的红外图像非均匀校正算法的基本原理和处理步骤。分析了运动目标对算法的影响,采用临近帧补全运动目标遮挡的场景的方法解决含运动目标的红外图像非均匀校正问题,通过模拟场景和实测场景验证了算法的有效性。最后详细分析了运动形变目标的形变方式、方向和范围等因素,并给出了形变目标的判定准则,进而给出了形变目标的帧间替换方法,提出了含形变目标的红外图像非均匀校正算法,通过模拟的红外序列图像校正实验验证了算法的有效性。
[Abstract]:The inhomogeneity of infrared image is accompanied by infrared focal plane array (Infrared Focal Plane Arrays,IFPA). The inhomogeneity in infrared image appears as fixed pattern noise, which seriously limits the imaging quality of infrared imaging system. At present, there are mainly black-based calibration methods and scene correction methods in the international non-uniformity correction of infrared images. This paper first introduces the composition of infrared imaging system and the development of infrared detectors, and introduces the classification and development of infrared image nonuniformity correction methods. Then the definition and source of nonuniformity and two response models of infrared detector are introduced. At the same time, several typical nonuniformity correction algorithms are introduced, and their basic principles, processing steps, advantages and disadvantages are analyzed. Then, the basic principle, processing steps, advantages and disadvantages of various image registration methods are introduced, and the basic principle and processing steps of infrared image non-uniform correction algorithm based on image registration are introduced in detail. The influence of moving target on the algorithm is analyzed. The method of cloaking the scene of the whole moving object by near frame is used to solve the problem of non-uniform correction of infrared image with moving object. The effectiveness of the algorithm is verified by simulation scene and real scene. Finally, the deformation mode, direction and range of the moving target are analyzed in detail, and the criterion of the deformation target is given, and the method of inter-frame replacement of the deformed target is given. A non-uniform correction algorithm for infrared images with deformed targets is proposed. The effectiveness of the algorithm is verified by simulated infrared sequential image correction experiments.
【学位授予单位】：国防科学技术大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41;TN215

【参考文献】