基于改进混合高斯模型的铁轨异物入侵检测方法

Railway foreign body intrusion detection method based on improved mixed Gaussian model

针对混合高斯背景模型存在的背景信息易被污染和前景异物目标检测不全的问题,提出了一种基于改进混合高斯模型(GMM)的异物入侵检测算法。该方法结合混合高斯背景模型和小波变换原理对运动目标进行检测。利用小波变换消除设备和噪声导致的背景建模过程产生的干扰点;提出基于邻域平均算法的像素修正算法,进行混合高斯建模,得到最终的背景模型,并通过差分法检测出异物目标。通过现场实验验证,平均前景误检率室外轨道正常环境下降低了57.85%,雪天环境下降低了55.02%,室内环境下降低了59.73%。实验数据表明:所提方法相比传统混合高斯模型方法目标检测完整度较高,检测效果较好。

Aiming at the problem that the background information of mixed Gaussian background model is easily polluted and the target foreign object is incompletely detected,a foreign object intrusion detection algorithm based on improved mixed Gaussian model(GMM)is proposed.The method combines the Gaussian background model and the wavelet transform theory to detect moving target.Wavelet transform is used to eliminate the interference points generated by the background modeling process caused by equipment and noise.A pixel correction algorithm based on neighborhood averaging algorithm is proposed,and mixed Gaussian modeling is earried out,and finally back ground model is obtained.The foreign object target is detected by the difference method.Through field experiments verification,the average prospect false detection rate decreased by 57.85%in normal outdoor track environment,and 55.02%in the snow track environment,and 59.73%in the indoor environment.The experimental data show that the proposed method has better target detection and better detection effect than the traditional hybrid Gaussian model.