城市道路环境通常是大规模且复杂的，使得无人机采集图像中的影子和模糊等因素会导致道路边界的模糊或不清晰，使得精确提取变得困难。为此提出基于无人机遥感的城市道路边界高精度提取。利用无人机遥感设备采集城市道路遥感图像，对其实施色调-饱和度-亮度(Hue Saturation Value，HSV)色彩空间变换、反射率转换、图像配准和融合处理，去除图像噪声，提高其清晰度;利用简单线索化迭代聚类(Simple Linear Iterative Clustering，SLIC)超像素分割算法分割遥感图像中的道路和非道路像素区域;通过多元局部二值模式(Local Binary Patterns，LBP)特征提取方法实现城市道路边界的高精度提取。实验结果表明，所提方法在提高遥感图像清晰度和去除图像噪声干扰的同时，具有较高的道路边界提取精度和提取效率，边界提取所用时间仅为2。74ms。
Efficient Extraction and Simulation of Urban Road Boundary under UAV Remote Sensing
Generally,urban road environments are large-scale and complex.During the process of collecting ima-ges by drones,the factors such as shadows and blurring may cause the boundaries of roads blurry or unclear,making precise extraction difficult.Therefore,a high-precision extraction method for urban road boundaries based on UAV re-mote sensing was proposed.Firstly,UAV remote sensing equipment was used to collect the images of urban roads,and then hue-saturation-value(HSV)color space transformation,reflectance conversion,image registration,and fusion processing were carried out to remove image noise and improve clarity.Secondly,the simple linear iterative clustering(SLIC)superpixel segmentation algorithm was utilized to segment road and non-road pixel regions in the remote sensing image.Finally,the multivariate local binary patterns(LBP)feature extraction method was adopted to achieve high-precision extraction of urban road boundaries.Experimental results demonstrate that the proposed method has high accuracy and efficiency in road boundary extraction while improving the clarity of remote sensing images and re-moving noise interference,and the extraction time is only 2.74ms.
NETL
NSTL
万方数据
