欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
背景提取是视频监控中运动目标检测的关键环节。由于前景的存在以及背景的变化，使得背景提取较难处理。为了获得合理的背景模型以满足实际应用中实时可靠的要求，针对视频序列图像中背景建模的难点和实际应用中的问题，提出了先利用帧间差分得到帧差图像，再计算帧差图像阈值来建立初始背景，以及利用高斯分布“3δ原则”更新背景的思想及算法。利用此方法对已储存的视频数据进行处理，算法可以有效准确地提取出背景图像，满足视频监控所需的实用性和可靠性。
英文摘要：
Background extraction is key link of moving objects detection in video monitoring. Due to the existence of the foreground and change of background, make background extraction more difficult. In order to obtain reasonable background model to meet the require ments of real-time reliably in practical application, according to the difficulty in the background modeling and actual application prob- lem,put forward by using frame difference get frame differential image,and then calculate frame differential image threshold to establish initial background, and use Gaussian distribution three δ thought and algorithm to update background. By using this method to process al ready stored video data, the algorithm can effectively, accurately extract background image, meet the practicability and reliability of video monitoring.
同期刊论文项目
同项目期刊论文
Performence Comparisons Between Cellular-only and Cellular/WLAN Intagrated Systems Based on Analytic
Reconfiguring Three-Dimensional Processor Arrays for Fault-Tolerance: Hardness and Heuristic Algorit
Efficient Localization for Moble Sensor Networks Based on Constraint Rules Optimized Monte Carlo Met
期刊信息
位置：