摘要
研究城市轨道交通控制保护区无人机巡护方法,用以解决无人机飞行状态巡护采集图像中出现特征缝隙和重叠等问题,提高巡护效果。使用小波变换等技术预处理无人机采集的图像,通过自相关函数和Hessian矩阵提取预处理后图像的角点,采用SIFT算法实现角点匹配,在此基础上利用加权平均方法完成图像融合,使用图像差分法对融合后的图像实行目标识别,通过识别结果实现城市轨道交通控制保护区无人机巡护。经验证,该方法具有良好的图像处理能力,图像融合后未出现特征缝隙和特征重叠的现象,具有良好的识别能力,完成城市轨道交通控制保护区无人机巡护效果较好。
An UAV(unmanned aerial vehicle)patrol method in urban rail transit control and protection area is studied to eliminate the feature gaps and overlaps appearing in the images collected by UVA in flight patrol and improve the patrol effect.Wavelet transform is used to preprocess the images collected by UAV.The corner points of the preprocessed images are extracted by means of autocorrelation function and Hessian matrix.The SIFT algorithm is used to complete corner point matching.On this basis,the weighted average method is used to complete the image fusion.The fused images are subjected to object recognition implemented by the image difference method.The UAV patrol in urban rail transit control and protection area is fulfilled with the recognition results.It is verified that the proposed method is of good recognition ability and image processing ability,so the phenomena of the feature gaps and feature overlaps do not occur after image fusion.Therefore,the method has a good effect for UAV patrol in urban rail transit control and protection area.
作者
毕景佩
韩旭
岳会婷
孙媛
石运财
BI Jingpei;HAN Xu;YUE Huiting;SUN Yuan;SHI Yuncai(Changsha University of Science&Technology,Changsha 410114,China;Zhengzhou University,Zhengzhou 450001,China;Zhengzhou Metro Group Co.,Ltd.,Zhengzhou 450000,China;North China University of Water Resources and Electric Power,Zhengzhou 450000,China;South Surveying&Mapping Instrument Co.,Ltd.,Guangzhou 510000,China)
出处
《现代电子技术》
2021年第15期65-69,共5页
Modern Electronics Technique
关键词
无人机巡护
城市轨道交通
图像采集
角点匹配
图像融合
目标识别
UAV patrol
urban rail transit
image acquisition
corner point matching
image fusion
object recognition
