摘要
从分形布朗模型出发,通过对自然背景和人造目标图像不同的灰度相关性分析,提出一种新的分形特征提取算法,根据新算法得到的分形特征分布规律,进而形成一种简单有效的自然背景中人造目标的检测算法。实验结果说明了该算法的性能。
A new method is proposed for estimating the fractal features of infrared images including manmade objects in the nature background, in which the correlation of intensity is considered. By analyzing the properties of fractal dimensions and modelfit errors, an effective algorithm based on these properties is presented for detecting manmade objects in the nature background. And the experiments have shown the effectiveness and feasibility of this new method.
出处
《激光与红外》
CAS
CSCD
北大核心
2003年第5期374-376,共3页
Laser & Infrared
