基本信息:
- 专利标题： 一种改进的ORB特征匹配方法
- 专利标题（英）：An improved ORB feature matching method
- 申请号：CN201810880470.X 申请日：2018-08-04
- 公开(公告)号：CN109146861A 公开(公告)日：2019-01-04
- 发明人： 郑茜颖 , 杨炳坤 , 程树英 , 张彩霞 , 林培杰 , 何煌城 , 俞金玲 , 陈志聪
- 申请人： 福州大学
- 申请人地址： 福建省福州市闽侯县福州地区大学新区学园路2号
- 专利权人： 福州大学
- 当前专利权人： 福州大学
- 当前专利权人地址： 福建省福州市闽侯县福州地区大学新区学园路2号
- 代理机构： 福州元创专利商标代理有限公司
- 代理人： 蔡学俊
- 主分类号： G06T7/00
- IPC分类号： G06T7/00 ; G06T7/66 ; G06T7/30 ; G06K9/46
The invention relates to an improved ORB feature matching method, comprising the following steps: step S1: adopting an improved FAST 14-24 method for Preliminary extraction of corner points was carried out and corner points were obtained; step S2: according to the obtained corner point, adopting a Shi-Tomasi corner detection algorithm for feature point selection to obtain feature points; step 3: processing the feature point set by using the gray centroid method to determine the direction of the feature point; step 4: according to that feature point set, adopting a retinal-like descriptor extraction algorithm to obtain a feature descriptor; step S5: according to the obtained feature descriptor, a learning method being adopted to extract the position of the sample point pair with low correlation to obtain the optimized feature descriptor; 6: carrying out feature matching by use Hamming distance. Compared with the existing rBRIEF descriptor, the optimized feature descriptor obtained by the invention has better robustness and higher accuracy.
公开/授权文献:
- CN109146861B 一种改进的ORB特征匹配方法 公开/授权日：2021-09-28
IPC结构图谱:
| G | 物理 |
| --G06 | 计算；推算；计数 |
| ----G06T | 一般的图像数据处理或产生 |
| ------G06T7/00 | 图像分析，例如从位像到非位像 |