基于先验知识的小目标快速随机识别方法

文章提出了一种基于先验知识的快速随机识别方法,能够从有背景噪声的遥感图像中快速识别出小目标。此方法基于不变矩进行识别,在识别中采用最大类间方差法自动选取阈值,使用中值滤波进行噪声的抑制,并且采用一种基于先验知识的随机匹配算法以加快匹配速度。实验结果表明,这种方法有很快的速度和比较高的精度。

Recognition of Small Targets by Random Matching Algorithm Base on Prior Experience

A system which can recognize small targets out of complex environments independently of postion.Size and orientation is proposed.The system first selects thresholds automatically for the original image to get several binary im-ages.Each threshold may be different from others.The invariant moments are calculated to determine whether the binary image is the target concerned or not.A random matching algorithm is used to improve speed of processing.It has proved that this system can get high speed and accuracy appropriate.