摘要
景象匹配系统由于其精度高的特点,广泛应用于各种远程精确末制导系统。由于成像时的自然条件变化,使得基准图与实时图之间存在较大的对比度差别,增大了匹配难度。本文基于传统的特征提取,针对具有人工景物的郊外景象,利用Hough变换和数学形态学进行线特征的改进,从而提升了匹配质量。通过对实际情景图像的仿真分析,验证了该算法的有效性。
Because scene matching system has the characteristic of high precision, it is widely used in all kinds of long-range precision end guidance system. There are big contrast differences between reference image and real-time image due to changes of natural conditions, so it is more difficult to match scenes. Based on the traditional feature ex-traction, for suburban scene with artificial scene,linear feature is improved by Hough transform and mathematical mor-phology to improve the matching quality. The effectiveness of this algorithm are verified through the simulation analysis with real scene images.
出处
《激光与红外》
CAS
CSCD
北大核心
2017年第8期1046-1050,共5页
Laser & Infrared
关键词
景象匹配
特征提取
低对比度图像
数学形态学
scene matching
feature extraction
low contrast image
mathematical morphology
