基于最小一乘和混沌遗传算法检测红外小目标

提出了一种基于最小一乘估计和混沌遗传算法进行背景预测检测红外小目标的方法.在建立最小一乘准则背景预测模型的基础上,根据最小一乘估计的性质,利用混沌序列内在的伪随机性,将混沌引入到遗传算法得到混沌遗传优化算法,以此解决最小一乘估计中极值的选取问题.将原始图像与预测图像相减得到预测残差图像后,利用基于二维指数熵的图像阈值选取快速算法进行分割.给出了实验结果与分析,并与基于遗传算法的最小一乘预测、最小二乘背景预测的检测算法作了比较.实验结果表明,提出的算法具有更高的检测概率和更好的检测结果.

A Method of Small Target Detection in Infrared Image Sequences Based on the Least Absolute Deviation and Chaos-genetic Algorithms

A method of small target detection in infrared image sequences was proposed based on the least absolute deviation background predication and chaos-genetic algorithms.Prediction model of the background signal based on the least absolute deviation criterion was founded.Based on characters of the least absolute deviation estimation,the extreme value was extracted by the chaos-genetic algorithms,obtained by using chaotic variable in genetic algorithms.The estimated image subtracted from the source image gave the residual image.And,the fast threshold selection algorithm based on the two-dimensional exponent entropy was used to segment the residual image.The experimental results are given and analyzed.Compared with the method based on least squares estimation and the traditional genetic algorithms,the results show that approach can precisely detect the small infrared target and has better results.