摘要
在背景为强杂波环境并且系统噪声影响比较小时,采用传统概念上提取弱小目标点方法的探测系统会主要被背景中的突变点和起伏干扰;针对复杂背景下的弱小目标图像提出了一种基于特征提取和理解的用于经过欧氏距离与灰度级均连续化处理后中间图像的聚类面积决策方法;算法利用直方图修正方法的结合来实现弱小目标图像背景中大面积对象干扰的类相似化,运算精度和实时性都比较高;实验结果表明,该方法大大减少了提取目标点时被复杂背景假点干扰的机会,提高了小目标探测的准确性。
While the effect of system noise is not strong in the environment of heavy background clutter, detecting system using the method in traditional concept for cxtracting weak and small targets will mainly be disturbed by signal singularity included in complex background. This paper proposes a method of clustering and area decision based on feature extraction and apprehending for image including weak and small targets, and this method is fit for processed image after Euclidean distance and gray level becoming continuous. The algorithm achieves comparability of species of large area included in background of weak and small targets image by using histogram correcting to make the calculating accuracy and speed higher. The results of simulations show that the method greatly reduces disturbance in extracting targets in complex background and improves the accuracy in small targets detection.
出处
《计算机测量与控制》
CSCD
北大核心
2009年第2期410-413,共4页
Computer Measurement &Control
基金
国家自然科学基金资助项目(60575013)
关键词
聚类
弱目标
小目标
面积决策
形态学
clustering
dim target
small target
area decision
morphology