多能见度下多视角动态图像传感网融合算法研究

项目名称： 多能见度下多视角动态图像传感网融合算法研究

项目编号： No.61203224

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 自动化学科

项目作者： 刘刚

作者单位： 上海电力学院

项目金额： 23万元

中文摘要： 随着无线传感网技术的发展，图像传感器作为新型的监测手段越来越被广泛使用，图像融合是图像传感网应用的关键环节。其中，图像融合算法的稳健性、实时性以及低功耗问题是非常迫切需要解决的问题。现有的图像融合方法大都以多尺度分析和融合策略为研究对象，而融合系统的稳健性和实时性问题是非常迫切需要解决的问题，这限制了图像融合的应用。本项目重点研究图像融合系统的稳健性问题，辅助研究其实时性问题。首先对不同传感器类型的融合系统外界干扰方式进行分析，提出描述融合系统稳健性的评价方法，分析后小波域的层间以及层内的分布特征关联性特性，利用HMM的统计特性提出一种提高图像融合效果且能够改善融合稳健性的融合策略；而后，为了提高融合系统的实时性，采用扩展Kalman滤波器或粒子滤波器方法，考虑动态图像帧间相关信息，提出一种非线性后小波域HMM模型的序列图像融合方法。

中文关键词： 图像融合；多处度分析；多视角融合；统计信号处理；

英文摘要： With the development of wireless sensor network technology, the image sensor as a new means of monitoring is increasingly widely used, image fusion is a key part of the image sensor network applications. Among them, the robustness of image fusion algorithms, real-time and low power consumption are very urgent problems need to be address.The existing image fusion methods are based on the multi-resolution analysis and fusion strategy as the object of study, and fusion system robustness and real-time problems is very urgent problem, which limits the application in image fusion. This project focuses on the study of image fusion system robustness issues, assisted on the real time problem. First of all the different types of sensor fusion system interference means undertakes an analysis, put forward to describe the fusion system robustness evaluation method, after the analysis of wavelet domain layer and layer within the distribution feature correlation properties, using HMM statistical properties of the proposed an improved image fusion effect and can improve the robustness of the fusion fusion strategy; and then, in order to improve the integration of the real-time performance of the system, using extended Kalman filter or particle filter method, considering the dynamic image frame information, this paper presents a

英文关键词： Image Fusion；Multiscale Analysis；Multi-view Fusion；Statistical Signal Processing；