基于改进SIFT图像匹配的无人机高精度避障算法设计  

Design of High-precision Obstacle Avoidance Algorithm for UAV Based on Improved SIFT Image Matching

作　　者：肖英楠[1] 孙抒雨[1] XIAO Yingnan;SUN Shuyu(College of Engineering and Technical,Chengdu University of Technology,Leshan 614000,China)

机构地区：[1]成都理工大学工程技术学院,四川乐山614000

出　　处：《机械制造与自动化》2022年第1期237-240,共4页Machine Building & Automation

摘　　要：为提高无人机避障能力,提出基于改进SIFT图像匹配的无人机高精度避障算法。通过角点匹配和多分辨模式识别,实现对无人机高精度避障的图像显著特征点检测,通过对比梯度分析和参数融合识别,实现对无人机高精度避障图像信息加权融合处理,采用改进SIFT图像匹配方法,匹配无人机高精度避障地理空间网格,实现避障算法的优化设计。仿真结果表明,无人机避障的自适应规划能力和避障精度较高。To improve the obstacle avoidance ability of UAV,a high-precision obstacle avoidance algorithm based on improved SIFT image matching is proposed.By corner matching and multi-resolution pattern recognition,the salient features of UAV high-precision obstacle avoidance image are detected.The weighted fusion processing of UAV high-precision obstacle avoidance image information is achieved by comparing gradient analysis and parameter fusion recognition.The improved SIFT image matching method is used to match UAV high-precision obstacle avoidance geospatial grid,and the optimization design of obstacle avoidance algorithm is realized.The simulation results show that the UAV has high adaptive planning ability and high obstacle avoidance accuracy.

关 键 词：改进SIFT图像 角点匹配 无人机避障 自适应控制 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]