论文部分内容阅读
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。
Fixed-weight background prediction is a simple structure, high processing speed infrared weak small target detection algorithm. However, its application is limited by its less efficient filtering than adaptive weighting filters. In this paper, the algorithm of fixed-weight background prediction is studied in detail from the aspects of operation time and SNR. The effects of different size prediction windows and various weighting coefficient templates on the detection of small targets are compared and analyzed. The filter composed of the optimal prediction window and the optimal fixed weight coefficient template is determined through theoretical analysis and experiments, Improve and verify the rationality of the above analysis through computer simulation.