CAJ | 학술논문

传统的三帧差分运动目标检测算法易出现空洞及虚假边缘等现象。为了对其进行改进，提出了一种融合边缘检测的三帧差分运动目标检测算法。首先对三帧连续图像采用 Canny 边缘检测算子快速提取边缘图像，然后对三帧连续的边缘图像进行改进的三帧差分运算，最后通过阈值分割和形态学处理完成对运动目标的提取。实验结果表明，该算法计算简单，连通性好且准确率高，可满足实时检测的要求。
전통적삼정차분운동목표검측산법역출현공동급허가변연등현상。위료대기진행개진，제출료일충융합변연검측적삼정차분운동목표검측산법。수선대삼정련속도상채용 Canny 변연검측산자쾌속제취변연도상，연후대삼정련속적변연도상진행개진적삼정차분운산，최후통과역치분할화형태학처리완성대운동목표적제취。실험결과표명，해산법계산간단，련통성호차준학솔고，가만족실시검측적요구。
The traditional three frame differential algorithm of moving target detection is easy to appear the hole edge and false edge. In order to improve it, this paper puts forward an algorithm of moving object detection based on the fusion of three frame difference and edge detection. The algorithm firstly uses Canny edge detection algorithm to extract edge image of three consecutive images, and then it improves the three continuous edge images of three frame difference algorithm, finally, by threshold segmentation and morphological, it completes the extraction of target. The method is simple in calculation and the experimental results show that the algorithm has high accuracy, good connectivity, and can meet the requirements of real-time detection.