| 改进帧差法和背景差法的多目标跟踪 |
| |
| 引用本文： | 高勇钢.改进帧差法和背景差法的多目标跟踪[J].巢湖师专学报,2013(6):14-16. |
| |
| 作者姓名： | 高勇钢 |
| |
| 作者单位： | 安徽师范大学数学计算机科学学院,安徽芜湖241000 |
| |
| 基金项目： | 安徽省高校自然科学基金项目（项目编号：KJ20122116） |
| |
| 摘 要： | 提出一种将改进的帧差法和改进高斯背景差法相结合来更好地分割出运动目标的算法，其中改进的帧差法是用连续3帧相互帧差之和可有效克服空洞效果；改进高斯背景差法是对像素点的邻域均值替代单个像素点建立高斯背景图像，这样可对光线变化有一定抗干扰能力。将二者结合起来，可较好地跟踪多个运动目标。
|
| 关 键 词： | 智能视频监控 高斯背景 多目标跟踪 连通域 形态学处理 |
MULTIPLE TARGET TRACKING ON IMPROVED FRAME DIFFERENCE AND BACKGROUND DIFFERENCE METHOD |
| |
| Authors: | GAO Yong-gang |
| |
| Institution: | GAO Yong-gang (College of Math and Computer, Anhui Normal University, Wuhu Anhui 241000) |
| |
| Abstract: | This paper proposes a new algorithm to better segment the moving target through combining an improved frame difference method and the improved Gaussian background difference method. For the improved frame difference method,three successive frame differences are used to effectively overcome the effect of hole, for the improved Gaussian background difference method,the pixel neighborhood averaging value replaces pixel dots to establish a single pixel Gaussian background image, so it has a certain anti-interference ability of the light changes. By combining the two method, it may be effective to track the multiple moving targets. |
| |
| Keywords: | intelligent video surveillance Gaussian baekground multi-target tracking connected component morphological processing |
| 本文献已被 维普 等数据库收录！ |
