论文部分内容阅读
介绍了一种红外图像背景抑制的非参数方法(E_kernel)。提出了一种弱小目标的管线检测算法。E_kernel方法不同于传统的线性或非线性背景预测,它对背景杂波分布的统计特性不敏感,受其影响较小,具有非参数特性。管线检测算法对序列图像做若干相同的顺序处理,采用并行分布式计算,处理时间短。仿真试验表明,该算法能有效地检测出低信噪比红外序列图像中的弱小目标的运动轨迹,具有较高的实时性。
A non-parametric method for background suppression of infrared images (E_kernel) is introduced. Proposed a weak target pipeline detection algorithm. E_kernel method is different from the traditional linear or nonlinear background prediction, it is not sensitive to the statistical characteristics of the background clutter distribution, less affected by it, with non-parametric characteristics. Pipeline detection algorithm to do some of the same sequential image sequence processing, the use of parallel distributed computing, processing time is short. The simulation results show that this algorithm can effectively detect the trajectory of weak targets in low SNR images and has high real-time performance.