基于形态滤波与遗传算法的目标检测技术  

Target detection based on morphological filter and genetic algorithm

作　　者：熊飞[1] 舒金龙 朱振福[1] 李军伟[3] 

机构地区：[1]北京环境特性研究所,北京100854 [2]中国航天二院,北京100854 [3]北京理工大学信息科学技术学院,北京100081

出　　处：《系统工程与电子技术》2005年第9期1561-1563,1579,共4页Systems Engineering and Electronics

摘　　要：针对红外图像背景复杂及点目标容易被淹没的问题,提出一种采用形态滤波与遗传算法相结合的目标检测新算法。该方法利用遗传算法的全局寻优能力对形态滤波结构元进行优化,以增强形态滤波对目标的检测与识别能力。为克服算法运算量大,实时实现困难的缺点,同时开展了基于DSP的实时信号处理器的研制及相关实验研究。试验结果表明,该方法能实时有效检测信噪比为2左右的点目标。To counter the complex background of infrared image, point target is easy to lose, so a novel method for the target detection and recognition is proposed, namely the genetic training algorithm for morphological filters , The method uses genetic algorithm to optimize the structure cell of morphological filters to improve the target detection and recognition ability. In other hands, it is difficulty to realize the calculator in real time, so some researches in image processing with the DSP are performed, Experiment results indicate that the method can successfully detect the target with SNR equaling 2, and it excels in real-time tasks.

关 键 词：红外图像 点目标 形态滤波 遗传算法 数字信号处理(DSP) 

分 类 号：TN249.3[电子电信—物理电子学]