摘要
图像并行处理研究较多,但鲜有分块文献以及图像分块机制研究。针对高分辨率天文图像数据量大、特征提取速度缓慢的问题,研究了图像分块并行处理策略。通过对高分辨天文图像在不同分块分辨率下的特征提取时间进行对比分析,找到特征提取速度最快的分块分辨率。通过实验对比和分析得到分块策略,通过实验验证了方法的有效性。
Aiming at the problem of large-scale high-resolution astronomical image data and slow feature extraction speed, this paper adopts block parallel processing strategy for high-resolution astronomical images. In the aspect of image parallel processing, more scholars have conducted researches, but it is clear that the literature discusses how to divide the block and what kind of blocking mechanism is used to block the image. This paper compares and analyzes the feature extraction time of high resolution astronomical images at different block resolutions to find the corresponding block resolution with the fastest feature extraction speed. Through the experimental comparison and analysis, a block strategy was obtained and the effectiveness of the method was verified by experiments.
作者
夏永泉
孙静茹
WU Xin-wen
支俊
王兵
谢希望
XIA Yong-quan;SUN Jing-ru;ZHI Jun;WANG Bing;XIE Xi-wang(School of Computer and Communication Engineering,Zhengzhou University of Light Industry,Zhengzhou 450001,China;Faculty of Engineering & Information Technology,Griffith University,Brisbane 4111,Australia)
出处
《软件导刊》
2019年第9期199-201,206,共4页
Software Guide
基金
国家自然科学基金项目(81501547)
河南省科技攻关项目(172102410080)
关键词
高分辨率
天文图像
分块机制
并行
特征提取
high resolution
astronomical image
segmentation mechanism
parallel
feature extraction
