| 基于CUDA的卫星影像快速Mask匀光方法研究 |
| |
| 引用本文： | 常 高,王华斌,李国元,刘金刚. 基于CUDA的卫星影像快速Mask匀光方法研究[J]. 科学技术与工程, 2014, 14(12) |
| |
| 作者姓名： | 常 高 王华斌 李国元 刘金刚 |
| |
| 作者单位： | 首都师范大学,国家测绘地理信息局卫星测绘应用中心,国家测绘地理信息局卫星测绘应用中心,中国科学院计算技术研究所 |
| |
| 基金项目： | 科技部科技支撑项目(2011BAB01B05)、国家高技术产业化项目 |
| |
| 摘 要： | 针对大数据量的卫星遥感影像进行匀光处理的过程计算量大,效率低的情况,本文提出了一种基于CUDA平台的卫星影像快速匀光并行处理方法,根据现有算法结构上的特点,将处理过程中的主要运算部分交由GPU完成。实验通过对资源三号卫星所得影像做匀光处理,对比了GPU与多核CPU运行速度和效果,分析了该方案的加速性能和实用性。实验表明,在保持原处理质量的同时,该方法构建的并行处理模型有效地提高了原有匀光算法的处理速度。
|
| 关 键 词： | 遥感影像 Mask匀光法 GPU CUDA |
| 收稿时间： | 2013-11-07 |
| 修稿时间： | 2014-03-26 |
Research of Fast Mask Dodging for Satellite Image Based on CUDA |
| |
| Chang Gao,Wang Hua-bin,Li Guo-yuan and Liu Jin-gang. Research of Fast Mask Dodging for Satellite Image Based on CUDA[J]. Science Technology and Engineering, 2014, 14(12) |
| |
| Authors: | Chang Gao Wang Hua-bin Li Guo-yuan Liu Jin-gang |
| |
| Affiliation: | Satellite Surveying and Mapping Application Center, NASG,Satellite Surveying and Mapping Application Center, NASG,Institute Of Computing Technology Chinese Academy Of Science |
| |
| Abstract: | Due to large size and high definition, the process of removing uneven illumination of satellite remote sensing image is higher time-consuming and lower efficiency than before. Based on CUDA platform, this paper presents a parallel method of removing uneven illumination of satellite image. According to the feature of the given algorithm, GPU performed as the main computing unit. The experiment compares the processing speed with that of multi core CPU, analyses the performance of speeding up. The result demonstrates that with this parallel model built in the paper, by remaining processing quality, the original algorithm of removing uneven illumination has been speeded up effectively. |
| |
| Keywords: | remote sensing image Mask dodging GPU CUDA |
|
| 点击此处可从《科学技术与工程》浏览原始摘要信息 |
|
点击此处可从《科学技术与工程》下载全文 |
|
