基于知识的旋转海面小目标检测

Knowledge-based detection for small target in rotation sea background

针对海面小目标处于海空线这一先验知识,提出了基于小波变换和多方向Gabor滤波融合的方法,检测任意角度的海天线,确定小目标的潜在区域,减少搜索区域与虚警情况,并利用运动目标的相关性检测小目标。结果表明,该方法对于复杂的海面背景是有效的,能够很好地检测出目标点。

An image fusion method based on wavelet transform and multidirectional Gabor filter is presented, according to the previous knowledge that small target is near sea-sky line.The location of seasky division line at any angle is detected, the potential area of the small target is determined ,the search area and false alarm is reduced, and the small target is detected by use of correlation of moving target . The experimental results show that the method is useful for small target detection in the complicated sea background.