基于改进SURF算法的双目视觉火灾定位方法  

Binocular vision fire location method based on improved SURF algorithm

作　　者：欧阳继能 卜乐平[1] 王腾[1] 杨植凯 袁鹏[1] 

机构地区：[1]海军工程大学电气学院,湖北武汉430033

出　　处：《消防科学与技术》2017年第11期1613-1616,共4页Fire Science and Technology

摘　　要：提出了一种基于改进SURF算法的双目视觉火灾定位方法。在使用SURF算法提高特征提取速度的基础上,引用了RANSAC算法提高了匹配的正确率。实验证明,该算法减少了运行时间和特征点的误匹配率,具有较好的实时性与鲁棒性,对于大空间火灾定位具有一定的实用价值。A binocular vision fire location method based on im proved SURF algorithm was proposed. On the basis of improving the speed of feature extraction by using SURF algorithm, RANSAC algorithm was introduced to improve the matching accu racy. The experimental results demonstrate that the algorithm shortened the operation time and mismatching rate of feature points, has better real-time performance and robustness, and is practical valuable for large space fire location.

关 键 词：火灾定位 双目视觉 SURF算法 RANSAC算法 

分 类 号：X921[环境科学与工程—安全科学] TP391.4[自动化与计算机技术—计算机应用技术]