Kim Sungho
Yeungnam University, 280 Daehak-Ro, Gyeongsan, Gyeongbuk 712-749, Korea.
Sensors (Basel). 2015 Mar 25;15(4):7267-93. doi: 10.3390/s150407267.
This paper presents a method for detecting high-speed incoming targets by the fusion of spatial and temporal detectors to achieve a high detection rate for an active protection system (APS). The incoming targets have different image velocities according to the target-camera geometry. Therefore, single-target detector-based approaches, such as a 1D temporal filter, 2D spatial filter and 3D matched filter, cannot provide a high detection rate with moderate false alarms. The target speed variation was analyzed according to the incoming angle and target velocity. The speed of the distant target at the firing time is almost stationary and increases slowly. The speed varying targets are detected stably by fusing the spatial and temporal filters. The stationary target detector is activated by an almost zero temporal contrast filter (TCF) and identifies targets using a spatial filter called the modified mean subtraction filter (M-MSF). A small motion (sub-pixel velocity) target detector is activated by a small TCF value and finds targets using the same spatial filter. A large motion (pixel-velocity) target detector works when the TCF value is high. The final target detection is terminated by fusing the three detectors based on the threat priority. The experimental results of the various target sequences show that the proposed fusion-based target detector produces the highest detection rate with an acceptable false alarm rate.
本文提出了一种通过空间和时间探测器融合来检测高速来袭目标的方法，以实现主动防护系统（APS）的高检测率。根据目标与摄像机的几何关系，来袭目标具有不同的图像速度。因此，基于单目标探测器的方法，如一维时间滤波器、二维空间滤波器和三维匹配滤波器，无法在适度误报的情况下提供高检测率。根据入射角和目标速度分析了目标速度变化。射击时刻远处目标的速度几乎是静止的，且增加缓慢。通过融合空间和时间滤波器稳定地检测速度变化的目标。静止目标探测器由几乎为零的时间对比度滤波器（TCF）激活，并使用称为改进均值减法滤波器（M-MSF）的空间滤波器识别目标。小运动（亚像素速度）目标探测器由小的TCF值激活，并使用相同的空间滤波器找到目标。当TCF值较高时，大运动（像素速度）目标探测器工作。最终目标检测通过基于威胁优先级融合这三个探测器来终止。各种目标序列的实验结果表明，所提出的基于融合的目标探测器在可接受的误报率下产生了最高的检测率。
