项目名称: 无人机航空侦察图像镶嵌与目标定位关键技术研究
项目编号: No.61201434
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 电子学与信息系统
项目作者: 李良福
作者单位: 陕西师范大学
项目金额: 22万元
中文摘要: 无人机航空侦察图像镶嵌与目标定位技术是当前军事与国防领域的研究热点,尽管国内外在此相关领域取得了很多研究成果,但是在应用实践中,我们发现还存在若干需要解决的关键技术问题。本项目拟采用视觉注意机制,研究基于感兴趣区域定位的图像镶嵌方法,研究基于摄像机运动估计的图像匹配方法,根据上下文线索探测感兴趣区域,通过估计图像的整体运动约束匹配区域的搜索范围和方向,以提高图像匹配的准确度。研究基于多层次处理的全景图像无缝平滑技术,从摄像机成像和光照原理上解决无缝平滑问题,校正图像序列之间的全局亮度不平衡效应,并研究采用多尺度小波分析方法对过渡重叠区域进行局部融合处理。探索从全局优化补偿参数累积误差、高精度图像校正、全景图像中的目标点与大地坐标系变换的数学模型等多方面提高全屏图像定位精度的计算方法。通过本项目的研究 有望解决战场图像侦察等军用光电系统领域的多项技术需求,从而产生重大的经济价值和社会效益。
中文关键词: 无人机;图像镶嵌;目标定位;累积误差;无缝平滑
英文摘要: Image mosaic and object localization technology for UAV aerial reconnaissance are hot research topics in the current military and defense fields. Although there are many research achieves at home and abroad in related areas, in the application of practice, we found that there are still a number of key technologies problem. This project intends to use the visual attention mechanism, and researches image mosaic method based on interest region localization. We research the image matching method based on camera motion estimation, which detects regions of interest according to context clues, and constraint the search range and direction of matching region by estimating the whole motion of image, in order to improve the accuracy of image matching. We research panoramic image smoothing technique based on multi-level processing, and solve the seamless smooth problem by camera imaging and illumination principle, in order to correct the effects of the global brightness imbalance between the image sequences. The multi-scale wavelet analysis method is studied for the local fusion process of the transition overlap region. We explore the cumulative error compensation of parameters from the global optimization view, high-precision image correction, the mathematical model from the target point in the panoramic image to the worl
英文关键词: Unmanned Air Vehicle;Image mosaic;Object localization;Accumulative error;Seamless smoothness