CAJ | 학술논문

为了获得理想的视频运动目标检测结果，本文提出了一种标记分水岭和形态学相融全的视频运动目标检测算法。首先采用各向异性扩散模型对视频图像进行预处理，消除噪声对目标检测的干扰，然后差分算法提取运动目标的轮廓，并对提取的目标轮廓进行形态学操作，最后采用标记分水岭算法对运动目标进行标记和分割，并采用仿真实验测试算法的性能。仿真结果表明，本文算法较好防止了“过分割”缺陷的出现，可以准确从复杂背景检测到运动目标，不仅提高了视频运动目标的检测正确率，并且加快了视频的检测的速度。
위료획득이상적시빈운동목표검측결과，본문제출료일충표기분수령화형태학상융전적시빈운동목표검측산법。수선채용각향이성확산모형대시빈도상진행예처리，소제조성대목표검측적간우，연후차분산법제취운동목표적륜곽，병대제취적목표륜곽진행형태학조작，최후채용표기분수령산법대운동목표진행표기화분할，병채용방진실험측시산법적성능。방진결과표명，본문산법교호방지료“과분할”결함적출현，가이준학종복잡배경검측도운동목표，불부제고료시빈운동목표적검측정학솔，병차가쾌료시빈적검측적속도。
In order to obtain ideal detection results of moving object, a new video moving object detection algo-rithm based on labeling watershed and detection algorithm and morphology is proposed in this paper. Firstly, the im-age is pre-processed by diffusion model to eliminate noise interference on object detection, and the difference algo-rithm is used to extract outline of moving object, secondly, morphological operations is use to process the target out-line, finally, marked watershed algorithm is used to segment moving object and performance is tested by simulation experiments. The simulation results show that the proposed algorithm can prevent over segmentation problems and accurately detect moving object from complex background, not only improve the detection accuracy and accelerate the speed of moving object detection.