安全监控系统中井下危险区域运动目标检测和跟踪具有重要作用,通过提前预警并结合联动系统可有效防止危险事故发生。提出基于背景差法改进的在线调节学习速率混合高斯模型和结合颜色梯度以及数学形态学的背景建模方法分别对于光线缓慢变化和突变不同环境下的移动目标进行检测。实验证明新算法能够准确定位和跟踪运动目标。
Detecting and tracking moving object under the mine danger zones plays an important part in coal mine monitoring and control system. Using linkage system can pre-warn and prevent the danger accident. New methods are used to detect moving objects under different illumination environment. Improved online Gaussian mixture model combines background subtraction melhod is used to update the background image under gradual illumination changes environment and color gradient combines morphologic method to update the background image under sudden illumination changes environment. The experiment results show that new method can detect moving object quickly and exactly.