弱小目标检测预处理算法对比研究  

Contrastive study on preprocessing algorithm of small and dim target detection

作　　者：袁广林[1] 李从利[2] 韩裕生[3] 袁宏武[3] 姚翎[3] 

机构地区：[1]解放军炮兵学院计算机教研室,安徽合肥230031 [2]解放军炮兵学院信息工程教研室,安徽合肥230031 [3]解放军炮兵学院军用光电工程教研室

出　　处：《红外与激光工程》2006年第z4期207-212,共6页Infrared and Laser Engineering

摘　　要：针对弱小目标检测与处理问题,对比研究了帧间差分法、背景预测法和小波分析法三种典型的弱小目标检测预处理算法。理论分析和实验验证表明,背景预测法的综合性能较好,只是在时间复杂度上较差,如果硬件资源允许,则是首选的算法;小波分析法在检测精度和抑制噪声方面较差,但其运算速度较快,如果对检测精度要求不高,则是首选的算法;帧间差分法检测率较低,综合性能最差,较少使用。研究为弱小目标检测预处理算法的选择提供了依据,为弱小目标检测预处理算法的研究和设计提供了重要的参考,具有重要的现实意义.Aiming at the problem on preprocessing of small and dim targets detection,three classic preprocessing algorithms of dim small target detection,temporal difference,background prediction and wavelet analysis are contrastively studied.The theory analysis and test indicates that background prediction algorithm is better in general capability,but of a sort in time complexity;the algorithm based on wavelet analysis is better in computing speed,but of a sort in detection precision and anti-noise;temporal difference is inefficient in detection and lowest in general capability.The study offers evidence for selecting preprocessing algorithm of dim small target detection,and important reference for studying and designing of preprocessing algorithm of dim small target detection.

关 键 词：目标检测 帧检差分 背景预测 小波分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]