摘要
以有效识别运动目标图像关键帧为出发点,设计运动图像关键帧快速跟踪系统。采集运动目标图像,将运动图像模拟信号传输至视频采集卡上,经由高速A/D转换成数字视频序列后,构建运动目标的图像模板,采用运动图像关键帧跟踪算法获取运动图像关键帧;通过图像交叉熵差异度序列计算,将熵值最大的图像作为系统跟踪到的运动图像关键帧。实验结果表明,该系统的运动图像关键帧跟踪效果好、全面性高,且具备低冗余度和高查全率。
In order to effectively identify the key frame of moving object image,a fast tracking system for key frame of moving object image is designed.The moving object image is collected.The analog signal of the moving object image is transmited to the video capture card and converted into the digital video sequence by high⁃speed A/D to construct the image template of the moving object.The moving image key frame tracking algorithm is used to obtain the key frame of the moving object image.By calculating the cross entropy difference sequence of the image,the image with the largest entropy value is obtained and taken as the key frame of the moving object image tracked by the system.The experimental results show that the dsigned system has a good tracking effect,widely tracking coverage,low redundancy and high recall ratio.
作者
聂上伟
NIE Shangwei(Chengdu University of TCM,Chengdu 611137,China)
出处
《现代电子技术》
2021年第3期30-34,共5页
Modern Electronics Technique
关键词
运动图像
关键帧跟踪
快速跟踪系统
目标检测
聚类算法
交叉熵
阈值
moving image
key frame tracking
fast tracking system
target detection
clustering algorithm
cross entropy
threshold value