摘要
提出了一种序列图像中弱小运动目标空时增强检测技术并进行了计算机仿真研究。目标被假设为在单帧图像上具有一定尺寸,利用二维空域模板对目标能量进行增强并作二值化处理。根据目标最大速度条件,在二值化后的图像序列中搜索可能目标轨迹,并与预定义的轨迹表(track table)进行匹配从而获得检测结果。该技术最大的优势在于同时在空域和时域对目标作增强处理,大大改善了信噪比,十分有利于实时实现。理论分析和大量仿真实验表明该技术在5帧图像集成时能稳定检测SCNR=2dB、大小超过3×3的弱小运动目标。
A technique of spatial-temporal enhanced detection for moving dim small target m image sequences is presented and relevant computer simulations are investigated. The assumption that a target occupies some pixels in a frame is adopted. A two-dimensional spatial template is used for enhancing target energy, and then the image is binarized according to a threshold. In the light of maximum velocity limits, all possible trajectories are searched in binaries sequences. And the final results are obtained by matching the trajectories with a predefined track table. The advantage of that technique is enhancing signal in spatial and temporal domain together, which leads to promotion of SNR and implementation in real time available. Theoretical analysis and many simulations show that a 3 3 or bigger target whose SCNR is 2dB can be detected steadily by integrating 5 frames.
出处
《电子测量与仪器学报》
CSCD
2006年第6期16-20,共5页
Journal of Electronic Measurement and Instrumentation
基金
国家"863"高技术计划(编号:2004AA823120)
国家自然科学基金(编号:10376005)联合资助项目
关键词
弱小运动目标
空时增强检测
轨迹表
计算机仿真
moving dim small target, spatial-temporal enhanced detection, track table, computer simulation.
