CAJ | 학술논문

针对视频图像运动目标检测算法数据量多、处理速度慢的缺点，提出了对运动目标检测算法的优化方法，使用基于TMS320C64X系列DSP的程序优化方法对运动目标检测代码进行优化，采用了编译器选项、内联函数与宏定义、查找表、软件流水等优化方法，最后综合分析了相关优化方法对于时间性能的改进情况，实验结果证明了优化方法的有效性。
침대시빈도상운동목표검측산법수거량다、처리속도만적결점，제출료대운동목표검측산법적우화방법，사용기우TMS320C64X계렬DSP적정서우화방법대운동목표검측대마진행우화，채용료편역기선항、내련함수여굉정의、사조표、연건류수등우화방법，최후종합분석료상관우화방법대우시간성능적개진정황，실험결과증명료우화방법적유효성。
Based on the characteristics of a lot image data and low processing speed in video-based moving object detection methods, some programming optimization skills for fast algorithms are summarized on TMS320C64x platform, including using compiler options, using intrinsic and macro definition, using look-up table, using loop optimizing, etc. Then the time per-formance of related optimization skills is analysed. Experimental results indicate the validation of the optimization methods.