摘要
本文根据遥感图像的目标识别要求,从舰船的目标识别要点出发,首先介绍舰船遥感图像的识别流程,针对舰船遥感图像在摄取中存在一定云块和复杂海况的噪声干扰问题,提出一种基于多级视觉感知算法去除海面背景干扰,获取舰船目标;考虑到获取到的舰船目标存在一定的虚假识别,本文引入基于SVM分类器方法有效的过滤掉虚假舰船目标,并在最后对舰船目标的分类识别结果进行了分析。
In this paper,the identification process of the ship remote sensing image has been first introduced into highlighting from the ship＇s target identification,which has been based on the requirements of remote sensing image target recognition. And a multi-level visual perception algorithm to remove sea background interference has been proposed to get the ship target. Taking into account there are some false identification in acquiring the ship target,a method based on SVM classifier which can effectively filter out false ship target has been introduced. In the end,we have been analyzed the results of the classification.
出处
《舰船科学技术》
北大核心
2014年第12期86-90,共5页
Ship Science and Technology
关键词
目标识别
视觉感知
SVM
target recognition
visual perception
SVM
