摘要
针对传统三帧差分法中目标检测不完整、自适应性差的问题,本文对此进行了改进。通过将三帧差分法与边缘检测算子相结合解决了目标边缘不连续性的问题。采用自适应阈值法对边缘检测算子的高、低阈值进行自动选取,从而提高了算法的自适应能力。采用数学形态学算法处理,填充了运动目标的部分空洞。实验结果表明改进算法能更清晰完整地检测出运动车辆目标。
In view of the problem of incomplete target detection and poor adaptability in the traditional three-frame difference method,this paper improves this.By combining the three-frame difference method with the edge detection operator,the problem of target edge discontinuity is solved.The adaptive threshold method is used to automatically select the high and low thresholds of the edge detection operator,thereby improving the adaptive ability of the algorithm.It is processed by mathematical morphology algorithm and fills some holes of the moving target.Experimental results show that the improved algorithm can detect moving vehicle targets more clearly and completely.
出处
《科技风》
2020年第29期69-70,共2页
基金
辽宁省自然科学基金项目(2015020061)。
关键词
运动车辆目标检测
三帧差分法
边缘检测
自适应阈值
数学形态学
Moving vehicle target detection
three-frame difference method
edge detection
adaptive threshold
mathematical morphology
