摘要
快速准确地定位无人机遥感图像中的光学标识是实现无人机实时操控的一个重要前提条件。针对光学标识的特征和形成机理,通过预先提取高亮区域,然后利用改进的模板匹配方法实现了高精度的光学标识识别,并通过充分利用累加计算,为匹配图像中光学标识所需要的加法次数不超过像素点数的35倍。不同环境下无人机拍摄图像数据的检测结果表明,算法的实时性好,且准确检测率可达到98%以上。
Fast and precise positioning of the optical mark in the remote image from the unmanned aerial vehicle （UAV） is an important precondition for UAV remote control. According to the characteristics and generation mechanism of the optical mark, an im- proved mask matching method is proposed to achieve high precision detection after the highlight area is obtained. Based on the matrices＇ accumulation operation,the needed number of the add operations to detect the optical mark is less than 35 times the number of the pixel numbers. Experimental results on various image data obtained by UAV demonstrate real-time detection can be achieved by the proposed method and the accurate detection rate is higher than 98%.
出处
《无线电工程》
2013年第10期51-54,共4页
Radio Engineering
关键词
模板匹配
累加
光学标识检测
无人机遥控
words mask matching
accumulation
optical mark detection
UAV remote control
