显著性的无人机远距离障碍物检测
姚俊; 蒋超; 华春生; 宋大雷
刊名沈阳理工大学学报
2016
卷号35期号:4页码:21-26
关键词形态Haar小波 障碍物检测 HOG特征 SVM分类器
ISSN号1003-1251
其他题名Remote Obstacle Detection Based on Saliency Feature for UAV
通讯作者姚俊
产权排序2
中文摘要针对旋翼无人机的避障问题,结合人眼视觉的显著性注意机制和HOG特征,提出一种实时的障碍物检测算法。该算法首先采用形态Haar小波分解的方法构建增强图像,然后根据增强图像的投影曲线提取出障碍物的候选区域。其后,对障碍物候选区域进行尺度归一化后提取HOG特征,并采用线性SVM分类器进行分类判别。该算法采用C++语言实现,在Celeron 2.3GHz处理器、2G内存的单板计算机上测试分辨率为640×480的VGA视频图像的处理速度约为14f/s。实验结果表明该算法满足无人机在低空环境下的障碍物的实时检测需求。
英文摘要In view of the rotor-wing UAV's obstacle avoidance problem,a real-time obstacle detection algorithm is presented,which is based on the significant of the human visual attention mechanism and HOG features. The proposed algorithm uses the method of two layers morphological Haar wavelet decomposition to built up the enhanced image. Then according to the projection curve of enhanced image,the border of obstacle candidate areas are extracted. Afterwards,the size of the obstacle candidate area is normalized,and the area's HOG feature is extracted,then the linear SVM classifier is used for classification. The method is implemented using C++ language For a resolution of 640×480 VGA video image,the test processing speed is about 14f/s on the single board computer with 2G RAM and Celeron 2.3GHz processor. Experimental results show that the algorithm satisfies the demand of real- time obstacles detection when UAV is flying at low altitude.
语种中文
内容类型期刊论文
源URL[http://ir.sia.cn/handle/173321/19226]  
专题沈阳自动化研究所_机器人学研究室
推荐引用方式
GB/T 7714
姚俊,蒋超,华春生,等. 显著性的无人机远距离障碍物检测[J]. 沈阳理工大学学报,2016,35(4):21-26.
APA 姚俊,蒋超,华春生,&宋大雷.(2016).显著性的无人机远距离障碍物检测.沈阳理工大学学报,35(4),21-26.
MLA 姚俊,et al."显著性的无人机远距离障碍物检测".沈阳理工大学学报 35.4(2016):21-26.
个性服务
查看访问统计
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。


©版权所有 ©2017 CSpace - Powered by CSpace