首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于CPU与GPU/CUDA的数字图像处理程序的性能比较
引用本文:岳俊,邹进贵,何豫航.基于CPU与GPU/CUDA的数字图像处理程序的性能比较[J].地理空间信息,2012,10(4):45-47.
作者姓名:岳俊  邹进贵  何豫航
作者单位:武汉人学洲绘学院,湖北武汉,430079
基金项目:国家自然科学基金资助项目(41074025)
摘    要:测绘系统通常有大量图像处理工作,而户外图像采集系统更需要极高的实时性。通过对GPU与CPU分别进行数字图像处理性能测试,结果表明GPU并行计算可以大幅提高图像处理性能:可优化图像预处理、后处理速度,使测绘系统更加实时高效。

关 键 词:CPU  GPU  CUDA  并行计算  数字图像处理

Performance Comparison of Digital Image Processing Based on CPU and GPU/CUDA
Abstract:Surveying and Mapping System usually has a lot of images to process and Out-door Image Collection System calls for high realtime performance. The result of digital image processing tests on GPU and CPU showed that the performance of digital image processing could be greatly improved by GPU-based parallel computing. GPUbased parallel computing can increase the speed of image pre-processing and image post-processing so that the Surveying and Mapping System becomes more efficient and has high real-time performance.
Keywords:CPU  GPU  CUDA  parallel computing  digital image processing
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号