|
| 视频媒体网络中基于轨迹优化的监控图像聚类算法 |
| Monitoring Image Clustering Algorithm Based on Trajectory Optimization in Video Media Network |
| |
| DOI：10.16018/j.cnki.cn32-1650/n.202304004 |
| 中文关键词: 视频媒体 视频网络 轨迹优化 监控图像 图像聚类 |
| 英文关键词: video media video network trajectory optimization monitoring image image clustering |
| 基金项目: |
|
| 摘要点击次数: 92 |
| 全文下载次数: 105 |
| 中文摘要: |
| 为了解决视频媒体网络环境中监控图像随机性较强、特征较为复杂，导致聚类效果不佳、空间畸变率高等问题，引入轨迹优化技术对视频媒体网络中监控图像进行聚类。通过采集视频媒体网络中监控图像的关键帧，挖掘监控图像的时空轨迹，并根据监控视频图像背景的隶属度计算结果，提取监控图像色彩、边缘等基本特征；计算监控内容、时间及空间的相似度，采用层次聚类算法对相似度较高的监控图像进行单元合并，并以视频轨迹为标准实现聚类融合，输出最终的聚类结果。实验结果表明，使用本文算法得出的聚类结果，平均畸变率仅为0. 85%，比传统算法降低了4. 65%。 |
| 英文摘要: |
| In order to solve the problems of strong randomness and complex features of monitoring images in video media network environment, which lead to poor clustering effect and high spatial distortion rate, trajectory optimization technology is introduced to cluster monitoring images in video media network. By collecting key frames of monitoring images in video media networks, mining the spatiotemporal trajectory of the surveillance images, and calculating the membership degree of the background of the monitoring video images, basic features such as color and edges are extracted from the surveillance images. The similarity of monitoring content, time and space is calculated, the hierarchical clustering algorithm is used to merge the monitoring images with high similarity, and the video track is used as the standard to achieve cluster fusion, and the final clustering result is output. The experimental results show that the average distortion rate of the clustering results obtained by this algorithm is only 0. 85%, which is 4. 65% lower than that of the traditional algorithm. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |