红外小目标图像快速核稀疏表示理论与方法研究

项目名称： 红外小目标图像快速核稀疏表示理论与方法研究

项目编号： No.61301207

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 无线电电子学、电信技术

项目作者： 魏长安

作者单位： 哈尔滨工业大学

项目金额： 24万元

中文摘要： 红外小目标图像信杂比低、对比度差、目标尺寸小、缺乏形状与纹理信息，对小目标进行检测与跟踪一直是计算机视觉领域的热点与难点问题，很多学者对此展开研究，但进展缓慢，亟需引入新理论、新方法。稀疏表示能够用较小的数据量捕捉图像中感兴趣目标的信息，在图像分类、人脸识别等图像处理领域取得了较好的效果，尤其是最近提出的核稀疏表示理论，进一步解决了稀疏表示在处理图像非线性问题上存在的理论瓶颈问题，在红外小目标图像处理上具有理论优势。本项目重点开展快速核稀疏表示理论研究及其在红外小目标图像处理上的应用方法研究，突破红外小目标图像快速核稀疏表示、核稀疏字典生成两个关键科学问题，研究基于核稀疏表示的红外小目标检测与跟踪算法，建立核稀疏红外小目标图像处理方法体系。本项目研究不仅能够进一步完善核稀疏表示理论，而且为红外小目标图像处理提供新的理论依据和处理方法，有望提高处理能力。

中文关键词： 机载前视；核稀疏表示；红外目标检测；红外目标跟踪；粒子滤波

英文摘要： Detecting and tracking small target in infrared image/imagery is a very hot and difficult issue in the field of Computer Vision due to the low signal-to-clutter ratio, low contrast of infrared image, and the small size, lack of shape and texture information of the target. There's little breakthrough on this issue recently, though many scholars research on it, so there is an urgent need to introduce new theories and methods to solve this problem. Sparse Representation (SR) is the theory that can represent the interesting of target in an image with a smaller amount of data, and the methods based on SR has achieved good results in image classification, face recognition etcetera. Especially, Kernel Sparse Representation (KSR) is presented to solve the nonlinear image processing problem, and outperforms SR in theories and image processing application. This project is to implement the theory research on KSR and application research on infrared small target image processing, and to solve two key science problems: fast KSR of infrared small target image and KSR Dictionary Learning, to improve the KSR-Based small target detecting and tracking algorithm, and to construct a method system of the KSR-Based infrared small target image processing. This research will not only further improve the KSR theory, but also provide hig

英文关键词： FLIR；kernel sparse representation；infrared small target detection；infrared small target tracking；particle filter