Design Of Anti-Collision And Localization System On Container Cranes

Container transportation has gradually become a major modern transport method.If the safe and efficient operating level is ensured,the throughput of the port will increase greatly.Since the spreader and container may have not enough height or drivers do not have enough view,collision of containers would still occur,bringing huge economic losses and even casualties.Meanwhile,the drivers will spend a lot of time in the link of containers' localization.It would reduce the operating efficiency and lower the speed of ports' throughput speed.Relying on scientific research project,this paper introduces the status of the application of computer-aided systems both in China and abroad.It compares the deficiency in using GPS and encoder to realize the function of anti-collision.It analyses the applying technical difficulties using lidar and vidicon at ports.Based on C#and the Open Source Computer Vision Library,it uses the shape characteristics of containers and video image processing to design the anti-collision and localization system on container cranes based on edge detection algorithm.As for the issue of contains' backward anti-collision,it compares two ways of raw data collecting using lidar and vidicon.It designs the system of contains'backward anti-collision based on lidar and the system of contains' backward anti-collision based on video image processing.It analyses the pros and cons of these two systems.Using vidicon can get more direct and more abundant information of contains' stacking.It can solve the issue of contains' collision in different rows.As for the issue of contains' localization,it analyses the geometrical characteristics of contains.It describes the different contain images in different directions.It raises the algorithm of contains' identification based on edge detection.It uses perspective transformation to calculate the relative position of the sling and contains.It generates the virtual reality image of contains' localization and provides direct information to drivers to help them in contains' localization.The major contributions of this thesis are as follows:1)It designs the system of contains' backward anti-collision based on lidar,by median filter and clustering,to detect edges and top surfaces of contains.2)It raises the algorithm of container identification based on video image processing.It analyses geometrical characteristics of contains and designs the system of contains' backward anti-collision based on vidicon.It solves the issue that the lidar can only collect the data of one section.3)It designs the system of contains' backward anti-collision based on video image processing,using C#and the Open Source Computer Vision Library.It generates the virtual reality image of contains' localization and provides direct information to drivers to help them in contains' localization.