杂乱
人工智能
转化(遗传学)
核(代数)
计算机视觉
计算机科学
滤波器(信号处理)
集合(抽象数据类型)
对比度(视觉)
目标检测
模式识别(心理学)
算法
数学
雷达
电信
化学
生物化学
程序设计语言
基因
组合数学
作者
Hu Zhu,Jieke Zhang,Guoxia Xu,Lizhen Deng
标识
DOI:10.1109/taes.2020.2982347
摘要
Detecting small targets in a complex background is always carried out by suppressing the background. The top-hat transformation is mainly utilized for background suppression in target detection. Many modified top-hat transformation methods are based on the different structures of the structural elements. However, there are two limitations. One is that the structural elements cannot sufficiently consider the contrast information between the target and surrounding area to enhance the target. Another is that structural elements should be set in advance and cannot adaptively suppress complex backgrounds. In this article, our proposed top-hat transformation is designed from two cases. First, an adaptive structural element based on a guided filter kernel is proposed for capturing the local features in infrared images for background suppression. Second, a balanced ring shape is used for two structural elements of top-hat transformation, which can utilize the contrast information between the target and background for target enhancement. More than 500 infrared target images are used in our experiment. The experimental results show that our algorithm achieves better performance in signal-to-clutter ratio gain, background suppression factor, and detection accuracy when compared with recent popular baseline methods.
科研通智能强力驱动
Strongly Powered by AbleSci AI