CAJ | 학술논문

为了增强红外或者可见光图像数据中的弱小目标检测,提出了一种采用模糊C均值(FCM)聚类与迭代最小二乘(RLS)自适应滤波相结合的背景抑制方法.假设待检测目标在图像帧上具有极小的空域扩展度,且受到强背景杂波的干扰.对输入的图像首先采用FCM聚类划分为灰度准半稳的子域,再将整帧图像均匀划分为相同的子块,然后在每个子块中针对每类子域利用RLS滤波估计背景杂波并另以去除,结果只剩下目标信号与残留噪声.大量仿真试验表明与其它传统方法相比具有更好的检测性能.
위료증강홍외혹자가견광도상수거중적약소목표검측,제출료일충채용모호C균치(FCM)취류여질대최소이승(RLS)자괄응려파상결합적배경억제방법.가설대검측목표재도상정상구유겁소적공역확전도,차수도강배경잡파적간우.대수입적도상수선채용FCM취류화분위회도준반은적자역,재장정정도상균균화분위상동적자괴,연후재매개자괴중침대매류자역이용RLS려파고계배경잡파병령이거제,결과지잉하목표신호여잔류조성.대량방진시험표명여기타전통방법상비구유경호적검측성능.
A method of background suppression using fuzzy c-means(FCM)clustering and recursive least square(RLS)filter is proposed to enhance the detection of dim small targets in IR or visual-light image data.The target to be detected is assumed to have a small spatial spread in a frame,and is obscured by heavy background clutter.The input data is firstly partitioned using FCM clustering,and each cluster is thought as a gray-level quasistationary subset.Secondly the image is partitioned to some sub-images uniformly,and then a RLS filter is applied to estimate background for each subset in each sub-image.Thus the background can be subtracted from input data,leaving components of the target signal in the residual noise.Many experiment results show better performance of detection by the method than by other traditional methods.