摘要：无人水面艇（USV）是近来新兴的海上作业平台,可以完成自主探测目标、躲避障碍、应急搜救等任务，而目标检测与跟踪系统是无人水面艇的关键技术,目标检测与跟踪可由雷达图像,红外图像，可见光图像完成，三者各有所长，目前国内的针对无人艇的可见光视觉系统研究尚不成熟。本文针对目前国内对这一系统的研究的不足之处和海面的实际情况研究了一套算法流程,并编写了相应的代码完成了一个目标检测与追踪系统，并用真实视频进行了测试。论文以无人水面艇的可见光视觉系统为研究对象,深入研究了目标分割，特征提取，图像识别，目标追踪。
论文开篇讨论了无人水面艇的现实意义及国内外的研究现状，其次针对目前的分割算法存在各种各样的不足找到一种较为适于水面情况的分割算法,并作出了改进。再次,根据一些论文的研究确定了适用于水面情况的纹理特征提取。再次,提出了针对样本数量较少情况下的图像识别算法。最后，选择了Camshift进行跟踪，并对Camshift作出了改进,使其能够追踪一些色彩暗淡的船只。最后用实地拍摄的视频验证,结果表明该系统能很好的完成识别与追踪。
关键词:无人水面艇；自适应局部二值化；灰度共生矩阵；自适应均值漂移
ABSTRACT：USV is a newly developed platform for offshore operation ,which can accomplish the tasks of autonomous detection,obstacle avoidance,search and rescue.Target detection and tracking system is the key technology of USV.We can detect target and track using radar images,infrared images,visible images,but the research of China on visible light vision system for USV is not mature,at present.According to the situation and sea environment, the paper do some research about algorithm.I made a system that can detect target and track,and use real video to test.The object of this paper is the visible vision system of USV,the key point is object segmentation,feature extraction,image recognition,target tracking.
The paper discuss the importance of USV and the research at home and abroad,and then find a method for the surface of sea that can avoid some common problems in segmentation of water surface images and make improvement, and then choose a way in some papers for texture feature extraction base on the situation of the surface of sea ,and then propose a way that don not need large sample for image recognition.Last,the paper choose Camshift for track and the method has been improved to enable tracking of ships which have faint color.Finally,I find that the system can identify and tack very well with video.
Key words: USV; Adaptive threshold; GLCM; Camshift
目录
第1章绪论 1
1.1无人艇简述 1
1.2无人艇的研究现状 1
1.3本文的主要研究内容 1
第2章系统的介绍 3
2.1运行环境 3
2.2采用的技术 3
2.2.1MFC简介 3
2.2.2OpenCV简介 3
2.2.3选用上述技术的理由 3
2.3视屏的获取及主要参数 4
2.4获取图像序列并处理的方式 4
2.5系统体系结构 4
2.6程序的主要流程 6
2.7程序的主要流程 6
2.8界面设计 7
2.9本章小结 无人水面艇目标检测与追踪系统研究与应用:http://www.youerw.com/jisuanji/lunwen_204697.html