摘要
固定权值背景预测是一种结构简单、处理速度较高的红外弱小目标检测算法。但因滤波效果不及自适应权值滤波器而限制了其应用。文章从算法的运算耗时和信噪比增益两方面,深入研究了固定权值背景预测算法。比较分析了不同大小预测窗口和各种权系数模板对小目标检测效果的影响,通过理论分析和实验确定了最佳预测窗口和最优固定权系数模板组成的滤波器,对常规的算法进行了改进.并通过计算机仿真验证了上述分析的合理性。
The algorithm of constant weight coefficient background forecast which has simple structure and high processing speed is used to detect weak and small IR targets.However,the algorithm is not widely used because of its effect is not as good as the algorithm of self-adaptive background forecast.The constant weight coefficient background forecast is studied from the aspect of processing time and the gain of SNR.The effect of weak and small IR targets detection is compared by using different size of prediction windows and three kinds of coefficient weight template.Based on the experimental results and theoretical analyze,the best size of prediction window and the optimal weight coefficient template of the filter are selected,for their contribution to the improvement of normal algorithm.Finally,all the analyze said above is proved by the computer simulation results.
出处
《红外与激光工程》
EI
CSCD
北大核心
2006年第z4期179-184,共6页
Infrared and Laser Engineering
关键词
背景预测
固定权值
小窗口
目标检测
Background prediction
Constant weight coefficient
Small window
Target detection
