基于DSP模式的计算机图像处理算法研究  

Research on Computer Image Processing Algorithms Based on DSP Mode

作　　者：刘蓓蕾 LIU Beilei(Xi'an Technology and Business College,Xi'an,Shanxi,710200,China)

机构地区：[1]西安工商学院,陕西西安710200

出　　处：《长江信息通信》2024年第9期65-67,共3页Changjiang Information & Communications

摘　　要：基于DSP架构模式,提出了一种CNN卷积神经网络算法,并将其运用到计算机图像处理中。研究过程中,采用DSP技术进行计算机图像获取、算法处理、算例分析和结果优化,大大提高了计算机图像算法处理质量和效率。经过算法测试验证,结果表明,基于DSP数字信号处理器搭建多DSP并行处理架构模式,采用CNN卷积神经网络算法进行计算机图像处理,能够提高图像处理精度。该算法运行时的性能较高,功能低,CPU占用率不高,且DSP计算机处理系统在多DSP并行处理架构模式下进行算法分析,系统的稳健性和可靠性高,能够适应不同规模级别下的计算机图像处理数据集的处理速度、精度、资源消耗和功率要求,可为计算机图像算法处理提供准确、高效、经济的解决方案,对于计算机图像处理算法设计和优化以及应用具有较好的实用参考价值。Based on the DSP architecture pattern,a CNN convolutional neural network algorithm is proposed and applied to computer image processing.During the research process,DSP technology was used for computer image acquisition,algorithm processing,case analysis,and result optimization,greatly improving the quality and efficicncy of computer image algorithm processing.After algorithm testing and verification,the results show that building a multi DSP parallel processing architecture based on DSP digital signal processors and using CNN convolutional neural network algorithm for computer image processing can improve image processing accuracy.This algorithm has high performance,low functionality,and low CPU usage during runtime.Moreover,the DSP computer processing system is analyzed in a multi DSP parallel processing architecture mode,and the system has high robustness and reliability.It can adapt to the processing speed,accuracy,resource consumption,and power requirements of computer image processing datasets at different scales.It can provide accurate,efficient,and cconomical solutions for computer image algorithm processing,and has good practical reference value for the design,optimization,and application of computer image processing algorithms.

关 键 词：DSP数字信号处理器 多DSP并行处理架构模式 CNN卷积神经网络算法 计算机图像处理方法 算法验证 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]