基于中值背景模型的运动目标自适应检测方法

目的 在保证准确性的前提下,降低运动车辆检测算法的计算量,加快处理速度,满足实时性要求,提出一种基于中值背景模型和自适应阈值的运动检测方法 .方法 基于当前帧与背景图像的差分图像,利用自适应阈值分别对差分图像的三个颜色通道进行二值化,从而实现运动目标的精确检测.同时,根据检测结果 ,采用中值更新策略实现背景图像的实时更新.结果 实验结果 表明,笔者算法可以从复杂交通场景图像序列中有效地检测出运动目标,并且算法计算量小,具有良好的鲁棒性与实时性.算法每帧处理时间比混合高斯降低43%,背景更新时间比一阶Kalman算法降低了45%.结论 算法能够很好地满足智能交通监控系统中运动车辆实时检测的要求.

Adaptive Motion Detection Based on Median Background Model

In order to decrease the computational cost and promote the processing speed and real-time performance of motion vehicle detection arithmetic with the same accuracy,an algorithm based on median background model and adaptive threshold is proposed.The difference image was got by the current frame subtracting from background image.Then the three color channels of the difference image are respectively binarized to accurately detect moving objects by adaptive threshold.According to the test results,median updating strategy is availed to achieve real-time background update.Experimental results on outdoor image sequences demonstrate that the proposed algorithm can effectively detect moving objects in complex traffic scenes.It has low computational cost,good robustness and real-time performance,and can meet the requirements of real-time detection of moving vehicles in intelligent transportation surveillance system.The processing time of proposed algorithm reduces 43% comparing with that of the GMM,furthermore,the processing time of the background updating reduces 45% comparing with that of the first order Kalman filter.