摘要
对海空背景下的电视和3-5μm红外目标提出了一种采用多向梯度表决融合的检测方法:依据 背景图像中点目标的奇异特性,得到图像的多向剃度检测结果,并进行表决融合。通过实测的电视和 红外图像仿真,结果表明,采用"and"判决规则的多向梯度表决融合的检测算法消除了云层、海浪、 海天线等背景干扰,在实现高检测概率的同时,可以达到较低的虚警概率。该算法特别适用于低帧率 的红外警戒系统,可检测信杂比为1的点目标。
A new detection algorithm based on multi-orientation gradient voting fusion for the TV and 3 infrared target under the sea background were proposed. Firstly, the target is detected through the multi-orientation gradient of image based on the target singularity. Then the result of the detection of TV and infrared image is fused through the voting theory. The computer simulation results show mat the algorithm based on multi-orientation gradient voting fusion is available and gives better result. The algorithm is suited especial for surveillance system of low frame rate and can detect the point targets with signal to clutter ratio equal to 1.
出处
《红外技术》
CSCD
北大核心
2005年第1期57-61,共5页
Infrared Technology
基金
武器装备预研重点基金(NO.614012)资助项目
关键词
双波段
杂波
点目标
目标检测
信杂比
虚警概率
天线
表决
新算法
背景图像
multi-degree multi-orientation gradient detection, voting fusion, target detecting, shipboard infrared search and track system
