CAJ | 학술논문

为加快图像预处理算法的执行速度，提出了基于计算统一设备架构（CUDA）的预处理算法来实现高速并行处理。分析了图像灰度化、高斯滤波以及直方图均衡化等预处理方法的原理，并对它们进行并行化分析，从而将 CUDA 并行计算技术引入到图像预处理算法。实验结果表明，此算法充分利用 GPU 的并行处理能力，与 CPU 串行处理方法相比，速度提高明显，有效提高数据处理能力。
위가쾌도상예처리산법적집행속도，제출료기우계산통일설비가구（CUDA）적예처리산법래실현고속병행처리。분석료도상회도화、고사려파이급직방도균형화등예처리방법적원리，병대타문진행병행화분석，종이장 CUDA 병행계산기술인입도도상예처리산법。실험결과표명，차산법충분이용 GPU 적병행처리능력，여 CPU 천행처리방법상비，속도제고명현，유효제고수거처리능력。
To speed up the execution speed of image preprocessing algorithm,the paper proposes preprocessing algorithm base on Compute Unified Device Architecture (CUDA )to achieve high speed parallel processing.It analyzes the principle of image gray,Gaussian filtering and histogram e-qualization preprocessing methods,and does paral-lel analysis,and then puts CUDA parallel compu-ting technology into the image preprocessing algo-rithm.Experimental results show that the image preprocessing algorithm based on CUDA takes full advantage of parallel processing capabilities of GPU,compared with CPU serial processing meth-od,significantly faster,improve data processing capability.