Abstract: Traditional angle measuring devices of bridge crane are installed in the crane system, usually using contact measuring device with potentiometer or photoelectric encoder, which have the disadvantages of complicated mechanical structure, poor independence and so on. Utilizing visual techniques, this paper proposes a computer vision based method to achieve the non-contact angle detection objective. Meanwhile, in order to improve the accuracy of payload identification and realize real-time operation, it uses the background modeling method with background subtraction for rapid recognition of moving payload to calculate the payload angle. Through experimental verification, it can be seen that this method recognizes moving payload sufficiently fast in real time, and calculate the load swing angle accurately. In particular, the algorithm does not depend on the geometry and quality distribution of the payload, which can be used to achieve accurate angle measurement for all types of payloads, making it suitable for industrial crane systems. 

Key words:  Bridge crane, background modeling; moving target detection, background subtraction, angle detection

摘要：

传统的桥式吊车摆角测量装置安装于吊车系统内部，多为采用电位器或光电编码器等方法进行检测的接触式测量装置，具有机械结构复杂、独立性差等缺点。算法将视觉引入桥式吊车摆角测量中，利用计算机视觉的方法实现负载摆角的非接触测量。同时，为了提高负载识别的准确性以及算法运行的实时性，利用背景建模结合背景减除的方法进行运动负载的快速识别。经过实验，算法能够准确、快速地识别运动负载，并对负载摆角进行准确计算，实验效果良好。特别的，算法无需考虑负载的几何形状和质量分布，对于各类负载均能实现准确的摆角测量，非常适用于实际工业吊车系统。

关键词:

桥式吊车, 背景建模, 运动目标检测, 背景减除, 摆角辨识