红外舰船小目标的检测

检测红外小目标常用的方法是先使若干帧图像叠加达到能量积累的效果,能量积累达到一定程度后,目标因能量较高而凸现出来.这种方法比较适合于对加性噪声不敏感或信噪比较高的情况.实验表明,当图像信噪比较低、对比度很差时不宜采用该方法.针对低信噪比小目标的检测问题,提出了两种方法：在天水线以下图像内相邻行、列分别差分结合Kapur最大熵分割目标的检测方法;在时域内用模板平滑结合Kapur最大熵分割目标的检测方法.实验证明,这两种方法简单易行,效果较好.

Detecting for infrared ship small targets

The general method of detecting the infrared small targets was that accumulating some frames to get enough energy of the small targets so as to make the targets appear. However it was suitable for the situations that there was mainly add-noise or SNR was high. Some experiments proved that this method was not effectual when there was very low SNR or bad contrast. Aiming at the problem of detecting the small target in low SNR, two methods to resolve this difficulty were given in this paper. First one was to difference the abutting rows or columns in the image under the horizon that was detected beforehand, followed by segmenting the image through the method of Kapur maximized entropy under horizon. The other one was to smooth the image in the time-fields combined with the segmenting method of Kapur maximized entropy. It is proved that the two methods in detecting infrared small targets are easy and effective.