摘要
针对天海背景远距离平视图像的成像特点,确定天海分界区域,可以减少目标分割时的计算量,抑制天海分界区域外不必要的噪声干扰。将求取天海分界方程归结为一个数理统计中的回归分析问题,通过将图像沿水平方向平均分成若干个区域,采用计算最大梯度得出每个区域的天海分界坐标,得到线性回归所需的区域坐标点集合,并用正态分布方法对野值进行剔除,运用最小二乘估计方法计算出天海分界线方程。结果表明,该方法可以准确地检测出天海分界线,且对目标背景特性不敏感。
According to the imaging characteristics of the long-range-observing image in the background of sea sky,it can reduce the amount of calculation of target segmentation and restrain the unnecessary noise interference in the boundary area of the sea if the sea-sky Boundary can be determined. In this paper,the sea -sky boundary equation was derived as a regression problem in mathematical statistics. The average image along the horizontal direction is divided into several regions, by dividing the maximum gradient that the sea-sky boundary coordinate calculation of each region,get a set of regional coordinates required for linear regression,using the least squares estimation method to calculate the sea boundary equation,and the outliers are eliminated by normal distribution method. The results show that the proposed method can detect the sea-sky boundary accurately and is insensitive to the characteristics of the target and the background.
作者
王学伟
WANG Xue-wei(Weifang University of Science and Technology,Weifang 262700,China)
出处
《火力与指挥控制》
CSCD
北大核心
2019年第5期82-84,89,共4页
Fire Control & Command Control
基金
佛山市科技创新平台基金资助项目(2017AG100132)
关键词
天海背景
天海分界
线性回归
最小二乘估计
sea-sky background
sea-sky boundary
linear regression
least squares estimation
