Feature Based Image Matching of NOAA/AVHRR Image
-
摘要: 为实现全球环境实时、动态监测的需要,选用Förstner特征匹配理论对NOAA/AVHRR影像进行匹配试验。Förstner特征提取之前,先根据熵的理论对影像进行预处理.特征提取工作量可减少到10%~20%。特征提取后的共轭影像,根据相似性尺度可建立一个可能的同名点匹配点表,再根据一致性尺度,采用Robust稳健估计剔除错匹配的点对,即可建立两幅影像的最终匹配点对。匹配结果表明,通过阈值的设置,可以在AVHRR多时相影像上提取足够数量的匹配点对,在行、列方向上平均误差约为0.6像素和0.3像素。Abstract: This poper is the preliminary research of "the hierachical method of global change monitoring". The theory of entropy has been used in FBIM-Feature Based Image Matching by the operator of Förstner. The preliminany system has been established for matching of NOAA/AVHRR image. The sufficient matching points can be collected using threshold. The mean error of row and line are 0. 6 pixel and 0. 3 pixel. The results of matcning are satisfactory.
-
Keywords:
- globle change monitoring /
- feature matching /
- entropy
-
计量
- 文章访问数:
- HTML全文浏览量:
- PDF下载量:
下载:
