红外图像的目标检测与识别技术研究.doc

红外图像的目标检测与识别技术研究
学号:
姓名:
指导老师:
摘要
低信噪比红外图像中目标的检测与识别技术是增加制导武器系统作用距离,加强系统防御能力的关键技术之一。本文研究了红外序列图像中目标的检测与识别技术,主要包括图像预处理、目标检测、目标识别。
本文首先分析了红外图像的特点,简要介绍了图像增强的一些方法。然后综合考虑系统的快速性及后续处理的需要,提出一种新的红外图像预处理方法。该算法充分利用红外图像中噪声的随机性和目标运动的相关性的特点,采用灰度对比值分割出潜在目标图像区域,再对潜在目标区域进行图像增强。
本文遵从最简单的DBT(先检测后跟踪)检测算法的设计思路,提出了一种基于帧间相关光流法的红外目标检测方法。实现了红外图像预处理与目标检测任务在一个完整的框架下的统一,实验结果证明了提出的目标检测方法兼顾了快速性和稳健性。
在分析红外目标图像的基本特征和不变性特征及典型的红外目标识别方法后,针对红外图像的特点提出了一组新的特征量,并结合局部最亮点轮廓线的不变矩对红外目标进行了分类识别。
关键词: 红外图像,目标检测,识别技术
ABSTRACT
The low signal-to-noise ratio in infrared image target detection and recognition technology is one of the key technologies that increase the distance of the role of guided weapons systems, system defense capabilities. This paper studies the infrared image sequences in target detection and recognition technologies, including image preprocessing, target detection, target recognition.
This paper first analyzes the characteristics of the infrared image, a brief introduction to some methods of image enhancement. And considering the need of fast and follow-up treatment, raised a new infrared image pre-processing methods. This algorithm takes advantage of the characteristics of the random noise in the infrared image and the correlation of the target motion, using the gradation ratio divided potential target image region, then the potential target area for image enhancement.
Design ideas in this ply with the most simple of the dbt (after the first detection, tracking) detection algorithm, proposed an inter-frame correlation-based optical flow method of infrared target detection method. Infrared image pre-processing and target detection task in the framework of plete unified, experimental results show that the proposed target detection method both fast and robust.
In the analysis of the basic features of the infrared target image and the invariant feature and typical infrared target recognition method for infrared image characteristics of a set amount of new features, combined with lo