CAJ | 학술논문

针对机器视觉系统中运动目标检测问题，在帧间差分法的基础上，提出了一种新的运动目标检测算法。采用改进的帧间差分法，融合分水岭算法分割特点，获得真实目标的轮廓，并进行实验验证。结果表明，该算法对帧差法进行改进后，得到的动态目标更接近于实际情况，改进的算法更适合于智能系统中消除杂波及对运动目标的检测。
침대궤기시각계통중운동목표검측문제，재정간차분법적기출상，제출료일충신적운동목표검측산법。채용개진적정간차분법，융합분수령산법분할특점，획득진실목표적륜곽，병진행실험험증。결과표명，해산법대정차법진행개진후，득도적동태목표경접근우실제정황，개진적산법경괄합우지능계통중소제잡파급대운동목표적검측。
Aiming at moving target detection in machine vision systems,an improved algorithm for moving target detec-tion based on frame difference algorithm is proposed.Combined with the segmentation characteristics of watershed al-gorithm,the real target profile can be obtained by using the improved frame difference method,and this method is veri-fied through the experiments.The results show that the dynamic targets obtained by this algorithm are closer to the ac-tual situation.In intelligent systems,the improved algorithm is more suitable to detect moving target and eliminate clutter.