Improved Moving Object Detection Algorithm Based on Gaussian Mixture Model
Lixia Xue · Video Engineering · 2012
An improved algorithm for moving objects detection based on Gaussian mixture model is proposed to make up the deficiencies of original method.Firstly,model updating mechanism is improved by using different learning rate in different period and updating the background model selectively.Secondly,improved Gaussian mixture model and frames substraction are combined by using twice and operation and once morphological dilation.Finally,the moving objects are got after the above-mentioned process.The experimental results indicate that the proposed method can effectively eliminate noise and shadow in a complex environment and improve the accuracy of moving objects detection.