摘要
目前的算法基本上都是针对单一图像源进行目标识别,方法的复杂度一般较高,运行速度慢,不能满足实际应用的需求。基于多源信息融合的水上桥梁目标识别方法将图像融合技术和目标识别技术有效结合起来,提出了一种新的目标识别方法。该方法主要分为3步:感兴趣目标区域提取,可疑桥梁目标提取,去除虚假目标。该方法综合利用了多源图像信息具有互补作用的特点,可准确地识别出目标。此外,该方法运行速度很快,具有在大图像上搜查可能存在目标的能力。
Most of current target recognition algorithms only utilize information from one kind of image to recognize targets. These algorithms always have high complexity, and cannot satisfy the application requirement. A target recognition algorithm based on the multi- source information fusion is proposed. The algorithm aims at recognizing bridges above the water. There are three steps in the algorithm. Firstly, regions of interest are abstracted. Secondly, dubious bridges are detected. Lastly, false hridges are deleted. The algorithm makes full use of different characters of different images, which can enhance recognition capability. Further, the algorithm runs rapidly, and have the capability to search targets in big images.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第22期5854-5856,5859,共4页
Computer Engineering and Design
关键词
目标识别
桥梁
多源信息融合
感兴趣区域
合成孔径雷达
target recognition
bridges
multi-source information fusion
region of interest
SAR （synthetic aperture radar）
