摘要
为了解决传统图像插值方法在处理效率上的问题,提出了基于小波双三次插值的遥感测绘图像空间分辨率检测方法。利用小波双三次插值将遥感测绘图像分割为空间单位。基于K-L变换原理构建图像滤波系数矩阵。采用异构并行体系,在CUDA架构下进行高密度值计算,实现了遥感测绘图像空间分辨率的快速检测。实验结果表明,本文方法能够在转换图像时保留更多的细节和纹理信息;在不同测绘比例下,能够在2.5 s内完成分辨率检测,且相对水平和垂直分辨率结果较高,适合人眼观察,具有实际应用效果。
In order to solve the problem of processing efficiency of traditional image interpolation methods,a spatial resolution detection method of remote sensing mapping image based on wavelet bicubic interpolation was proposed.Wavelet bicubic interpolation is used to segment remote sensing mapping image into spatial units.The image filtering coefficient matrix is constructed based on the principle of Kosambi-Karhunen-Loève theorem.Adopting a heterogeneous parallel system,high-density value calculation is carried out under the CUDA architecture,achieving rapid detection of spatial resolution in remote sensing surveying and mapping images.The experimental results show that the proposed method can preserve more details and texture information when converting images.Under different mapping ratios,resolution detection can be completed within 2.5 s,with relatively high horizontal and vertical resolution results,suitable for human eye observation and practical application effects.
作者
万计
WAN Ji(Hunan Quality Inspection Center for Surveying and Mapping Product,Changsha,Hunan 410007,China)
出处
《计算技术与自动化》
2024年第3期108-113,共6页
Computing Technology and Automation
关键词
小波双三次插值
遥感测绘图像
检测方法
空间分辨率
wavelet bicubic interpolation
remote sensing mapping image
detection method
spatial resolution
