| 栾庆磊,周希勇,赵为松,朱广,邓从龙.AKAZE与高斯混合模型相融合的运动目标检测算法[J].安徽建筑大学学报,2023,31(1):64-69，96 |
| AKAZE与高斯混合模型相融合的运动目标检测算法 |
| Moving Target Detection Algorithm Based on AKAZE and Gaussian Mixture Model |
| |
| DOI： |
| 中文关键词: 运动目标检测 高斯混合模型 去抖动 AKAZE |
| 英文关键词: moving object detection Gaussian mixture model removing jitter interference AKAZE |
| 基金项目:安徽省高校优秀青年人才支持计划项目（gxyqZD2018058）；安徽省教育厅科学研究一般项目（KJ2019JD25） |
|
| 摘要点击次数: 6392 |
| 全文下载次数: 0 |
| 中文摘要: |
| 传统高斯混合模型在抖动干扰下无法取得理想的运动目标检测效果，但实际的各种复杂情况下摄像系统均会存在抖动。针对这一问题，文中提出一种抖动干扰下运动目标检测算法：首先利用基于 AKAZE 特征点匹配算法，进行动态图像精确配准，消除图像抖动，做出运动补偿；其次利用高斯混合模型对配准后的视频序列进行运动目标检测，然后利用三帧差分技术对检测出的运动目标轮廓进一步补偿；最后使用连通域分析补全运动目标。实验表明，本文提出的方法能够满足检测要求，可以在抖动干扰下有效检测出运动目标。 |
| 英文摘要: |
| The conventional Gaussian mixture model cannot achieve the ideal moving target detection under jitter interference，but thecamera system is subject to jitter in various complex situations in practice. To solve this problem，a algorithm for moving target detectionunder jitter interference is proposed in this paper. Firstly，the AKAZE feature point matching algorithm is used for accurate dynamicimage registration to eliminate image jitter and make motion compensation. Secondly，the Gaussian mixture model is used to detect themoving target in the aligned video sequence，then the three-frame difference technology is used to further compensate the detectedmoving target contour. Finally，the connected domain analysis is used to complete the moving target. The experiments show that theproposed method is able to meet the detection requirements and effectively detect moving targets under jitter interference. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
