Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/10142
標題: 以GPU平行運算加速邊界偵測分析之研究
An Analysis on GPU-based Parallel Edge Detection
作者: 吳晨郡
Wu, Chen-Chun
關鍵字: 影像處理;image processing;邊界偵測;平行運算;GPU;CUDA;parallel programming;CUDA;edge detection;GPU.
出版社: 土木工程學系所
引用: 辛大紅,2008,基於CUDA的圖像邊緣檢測方法,杭州電子科技大學學報,Vol. 5,No. 6,pp. 163-166。 張舒、褚豔利、趙開勇、張鈺勃,2009,GPU高效能運算之CUDA,中國水利水電出版社。 劉海波、沈晶、郭聳,2010,Visual C++數字圖像處理技術詳解,機械工業出版社。 蔡榮得,2011,國立中興大學土木學系影像處理碩士班上課講義。 林俊淵、周嘉奕、林郁翔、李昇達、陳昱蓉、黃宣穎、李天齡,2011,CUDA 輕鬆上手─新世代 GPU 應用技術,松崗出版社。 許學貴、張清,2011,基於CUDA的高效并行遙感影像處理,地理空間信息,Vol. 9,No. 6,pp.47-55。 曾勝田、劉羽、馬夢琦,2011,基於CUDA的Prewitt算子并行實線,微計算機應用,Vol. 32,No. 11,pp.71-75。 ASUSTeK Computer Inc., 2012, http://tw.asus.com/Server _Workstation/Servers/ESC1000_Personal_SuperComputer/#overview, accessed on 26th Jun., 2012. Bradski, G. and A. Kaehler, 2008, Learning openCV, O’REILLY. Capcom Interactive, Inc., 2012, http://www.mysmurfsvillage. com, accessed on 25th May, 2012. Chen, P. C., 2008, http://www2.kimicat.com/%E6%94%B9 %E8%89%AF%E7%AC%AC%E4%B8%80%E5%80%8B%E7%A8%8B%E5%BC%8F, accessed on 24th Apr., 2012. Christophe, E., J. Michel and J. Inglada, 2011, Remote Sensing Processing: From Multicore to GPU, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 4, No. 3, PP. 643-652. CIM(Centre for Intelligent Machines), 2012, http://www.ci m.mcgill.ca/~image529/TA529/Image529_99/assignments/edge_detection/references/sobel.htm, accessed on 5th Jun, 2012. Gonzalez, R. C. and R. E. Woods, 2008, Digital Image Processing, 3rd ed., Pearson Prentice Hall. Intel Corporation, 2012, http://ark.intel.com/zh-tw/products/41 313/Intel-Xeon-Processor-W3530-(8M-Cache-2_80-GHz-4_80-GTs-Intel-QPI), accessed on 26th Jun., 2012. Marengoni, M. and D. Stringhini, 2011, High Level Computer Vision using OpenCV, IEEE International Conference on Graphics, Patterns and Images Tutorials (SIBGRAPI-T), pp. 11-24. Nickolls, J., I. Buck, M. Garland and K. Skadron, 2008, Scalable Parallel Programming with CUDA, Queue, Vol. 6, Issue 2, pp.40-53. NVIDIA, 2011, NVIDIA CUDA C Programming Guide, Version 4.1. NVIDIA, 2012, http://www.nvidia.com.tw/object/product_tesla _C2050_C2070_tw.html, accessed on 8th May, 2012. OpenCVWiki, 2012, http://opencv.willowgarage.com/wiki/ OpenCV_GPU?highlight=%28%28OpenCV+GPU+FAQ%29%29, accessed on 1st Jun., 2012. PCPOP, 2012, http://product.pcpop.com/000303060/Original Pic/004284371.html, accessed on 26th Jun., 2012. Pulli, K., A. Baksheev, K. Kornyakov and V. Eruhimov, 2012, Real-time Computer Vision with OpenCV, Magazine Communications of the ACM, Vol. 55, No. 6, pp. 61-69. Prewitt, J., 1970, Object Enhancement And Extraction, Picture Processing and Psychopictorics ( B. Lipkin and A. Rosenfeld, ed.), Academic Press. Roberts, L. G., 1963, Machine Perception of Three-Dimensional Solids, Massachusetts Institute of Technology, Lexington Lincoln Lab. Sanders, J. and E. Kandrot, 2010, CUDA By Example: an introduction to General-Purpose GPU Programming, Pearson Prentice Hall. Sobel, I., 1970, Camera Models and Machine Perception, PhD thesis, Department of Computer Science, Stanford University. Wadworth Center, 2012, http://www.wadsworth.org/spider_doc /web/docs/convolve.html, accessed on 6th Jun., 2012. Wang, M., and C. H. Lai, 2009, A concise introduction to image processing using C++, Chapman & Hall/CRC. Yang, Z., Y. Zh, and Y. Pu, 2008, Parallel Image Processing Based on CUDA, IEEE International Conference on Computer Science and Software Engineering, Wuhan, China, Vol. 3, pp. 198–201. Zhang, N., J. Wang and Y. Chen, 2010, Image parallel processing based on GPU, IEEE International Conference on Advanced Computer Control (ICACC), Shenyang, China, pp. 367–370. Zhao, F., 1998, Use of the Laplacian of Gaussian Operator in Prostate Ultrasound Image Processing, IEEE 20th International Conference on Engineering in Medicine and Biology Society, Hong Kong, China, Vol. 2, pp. 812-815.
摘要: 
在遙感探測的數位影像處理中,邊界偵測(Edge detection)是應用於影像分析與判釋的基礎技術,廣泛應用於目標跟蹤、影像壓縮和電腦視覺等領域。
隨著遙測影像資料的增加,邊界偵測的計算量變得越大,傳統串列運算系統受限於中央處理器(Central Processing Unit;CPU)的計算效率,耗費的時間過久。為了提高演算法的計算效率。本文利用NVIDIA公司開發的統一計算設備架構(Compute Unified Device Architecture;CUDATM)來執行基於圖形處理器(Graphics Processing Unit;GPU)的平行運算邊界偵測,將CUDA可以同時執行多個執行緒來進行大規模快速計算的特性,應用到數位影像處理,以解決了邊界偵測基於CPU 運算效率低的問題。
本文以影像處理中常用的邊界偵測為例,敘述利用CUDA的平行處理方法,並針對解析度不同的影像及邊界偵測演算法進行實驗,同時與串列演算法的處理時間進行比較,取得加速比。實驗成果顯示,利用CUDA的平行演算法在處理數位影像的邊界偵測計算中,加速效果十分顯著,與傳統CPU串列處理方法比較,效率可以提升約20到50倍,有效提高處理能力。

Edge detection is a fundamental technique in digital image processing and remote sensing, especially on image analysis and interpretation. It is also widely used in object tracking, image compression and computer vision.
With the emerging development of remote sensing technology and the increment of image data, the computation in edge detection is getting larger than ever. Due to intensive calculation in edge detection, traditional CPU-based algorithm is time-consuming. In order to improve the algorithm''s efficiency, this study presents a parallel algorithm for processing edge detection based on Graphic Processing Unit (GPU) which support NVIDIA''s Compute Unified Device Architecture (CUDA) to solve the low efficiency problem by multi-threading edge detection algorithm
This paper also makes analysis of the efficiency about the image size and edge detector''s mask size. Numerical experiments show that the speed of the developed GPU-based parallel algorithm can be improved by up 20-50 times compared with CPU-based pipeline algorithm.
URI: http://hdl.handle.net/11455/10142
其他識別: U0005-2108201218050800
Appears in Collections:土木工程學系所

Show full item record
 
TAIR Related Article

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.