摘要
城市高精地图是汽车无人驾驶技术的基础,现阶段高精地图车道采集多采用车载移动测量设备获取数据,该技术存在效率低、需要透视变换及成像范围小的不足。文章基于无人机摄影测量获取的影像,采用U-Net识别道路范围,通过HSL颜色变换和Sobel算子计算提取车道信息,利用滑动窗口进行数据检验,该方法可以实现城市高精度地图车道线的提取,且精度符合要求。
Urban high-precision maps are the foundation of autonomous driving technology for automobiles.Currently,high-precision map lane acquisition mostly uses on-board mobile measurement equipment to obtain data.This technology has shortcomings such as low efficiency,requiring perspective transformation,and small imaging range.The article is based on images obtained from drone photogrammetry,uses U-Net to identify the road range,extracts lane information through HSL color transformation and Sobel operator calculation,and uses sliding windows for data verification.This method can achieve high-precision extraction of lane lines in urban maps,and the accuracy meets the requirements.
出处
《智能城市》
2023年第7期33-35,共3页
Intelligent City