摘要
红外小目标检测一直是红外图像处理的难点之一,由于多种因素的影响,红外小目标容易被覆盖。分析红外图像特征,采用形态学对图像进行背景噪声抑制,在去除大部分噪声的情况下,首先利用灰度信息确定目标点的位置,然后利用区域梯度信息进而确定目标尺寸大小,对仿真图像进行处理并与K均值聚类法和形态学算法进行比较。实验结果表明:在低噪声情况下,三类算法均能有效地进行小目标检测,但在噪声复杂,信噪比较低的情况下,K均值聚类法未能检测出目标,形态学算法产生了多个虚警,而该算法依然能有效检测出小目标。
Due to the influence of the background and the other factors,infrared small target is easy to be covered,thus,infrared small target detection is always a difficulty in infrared image processing. An infrared small target detection method based on area gradient was proposed. Through analyzing the features of IR image,the background noise of the image was suppressed by morphology. The positions of target points were located by using gray value,and then the sizes of target points were decided by using area gradient information. The simulation images of different SNR were processed by this method,and this method is compared with K-means clustering and morphological algorithm. The results show that three methods can all effectively detection the target in the condition of high SNR,but only the proposed algorithm can detected the small target in low SNR image.
出处
《激光与红外》
CAS
CSCD
北大核心
2016年第12期1547-1550,共4页
Laser & Infrared
关键词
小目标
虚警
区域梯度
信噪比
small target
fare alarm
area gradient
SNR
