基于机器视觉的自动扶梯速度偏差及制停距离检测系统

Research on Escalator Speed Deviation and Stopping Distance Detection System Based on Machine Vision

针对自动扶梯梯级扶手带速度偏差、制停距离的测量问题,设计了一款采用机器视觉的自动扶梯速度偏差及制停距离检测系统。在对自动扶梯的实时运行状态进行检测前,需要输入待测自动扶梯的名义速度、提升高度、梯级宽度与高度、梯级制动载荷等性能参数;使用工业相机对梯级扶手带的速度进行监测,为了增强其自适应能力,采用了差图像速度测量算法与光流算法相融合的方法,经过大量样本训练出图像速度非零点的各个像素点,计算出非零点像素的光流值,易于提取复杂动态目标梯度变化的特征,建立图像动态识别测量模型,使其适用于较为复杂的动态测量对象。经验证表明,该系统在复杂环境下识别测量的梯级扶手带速度偏差和制停距离精确度高,将提高自动扶梯的隐患检出的能力,确保运行安全。

Aiming at the measurement problems of speed deviation and braking distance of escalator step handrail,a detection system for escalator speed deviation and braking distance using machine vision was designed.Before testing the real-time running state of the escalator,it was necessary to input the performance parameters such as the nominal speed,lifting height,step width and height,step braking load,etc.of the escalator to be tested.An industrial camera was used to monitor the speed of the step handrail.In order to enhance its self-adaptive ability,the method of combining the differential image velocity measurement algorithm and the optical flow algorithm was adopted.After a large number of samples were trained,each pixel of the image velocity was non-zero,and the optical flow value of the non-zero pixel was calculated,which was easy to extract and complex.The characteristics of the dynamic target gradient change were established,and the image dynamic recognition measurement model was established to make it suitable for more complex dynamic measurement objects.It verifies that the system has high accuracy in identifying the measured step handrail speed deviation and stopping distance in complex environments,which improves the ability to detect hidden dangers of escalators and ensure safe operation.