摘要
由于在影像拼接过程中存在清晰度低的问题,为此,提出基于改进SPHP(shape-preserving half-projective)算法的无人机遥感影像智能拼接方法。确定目标图像的极值点后,采用SIFT(scale invariant feature transform)作为特征参数,实现对无人机遥感影像特征的提取。在影像匹配阶段,对图像进行单应变换处理后,在待拼接无人机遥感影像特征重合的前提下,利用薄板样条函数(thin plate spline,TPS)中的径向基函数对图像进行变形处理,以此消除投影偏差,结合不同方向上的变形量,实现对无人机遥感影像的拼接。在测试结果中,拼接后影像的信息熵稳定在7.50以上,灰度方差乘积稳定在10.80以上。
Due to the problem of low definition in the process of image stitching,an intelligent stitching method for UAV remote sensing images based on an improved SPHP algorithm is proposed.After determining the extreme points of the target image,SIFI is used as the feature parameter to extract features from unmanned aerial vehicle remote sensing images.In the image matching stage,after performing homography transformation on the image,the radial basis function in the TPS function is used to deform the image under the premise of overlapping features of the unmanned aerial vehicle remote sensing image to eliminate projection deviation.Combining the deformation in different directions,the unmanned aerial vehicle remote sensing image is concatenated.In the test results,the information entropy of the spliced image is stable at more than 7.50,and the gray variance product is stable at more than 10.80.
作者
国仲凯
彭树鸿
郑福海
GUO Zhongkai;PENG Shuhong;ZHENG Fuhai(Inner Mongolia Surveying and Mapping Geographic Information Center,Hohhot 010010,China;The Second Geographic Information Cartography Institute of the Ministry of Natural Resources,Harbin 150000,China)
出处
《技术与市场》
2023年第9期24-27,共4页
Technology and Market
关键词
改进SPHP算法
无人机遥感影像
智能拼接
尺度空间
单应变换处理
径向基函数
投影偏差
improving the SPHP algorithm
unmanned aerial vehicle remote sensing image
intelligent splicing
scale space
monomorphic transformation processing
radial basis function
projection deviation
