Research On Anti-lifting Methods For Container Trucks Based On Video Image Processing

With the increasing demand for global container transportation,the operations of container yard at the terminals are becoming busier.During the operation process of lifting containers by the cranes,the current anti-lifting system employed on the container trucks,which is based on the laser range finder,only considers whether there is a gap between the container and the truck.It is lack of a comprehensive judgement on the abnormal conditions when the truck is lifted by the crane,with troubled by the fast wear and tear on devices and high cost.Hence,a new anti-lifting system for container truck based on the video image processing gains more and more attention and research.This paper mainly studies the anti-lifting methods for container trucks based on video image processing,including the core algorithm for the feature extraction and detection on the wheels of trucks as well as the completion of the whole anti-lifting system for container trucks.In the first place,concerning to the complex application condition,a feature extraction method of truck wheels based on the edge extraction and arc grouping is proposed.After the edge detection,the edge segmentation will be divided by combining corner detection with boundary tracking.Then,based on the shape feature of the wheels,two selection strategies on the effective arcs are provided.Finally,those effective arcs are obtained by excluding the invalid edge information from the edge segments.Secondly,since the background of the images obtained in this subject is too complicated,the simple contour fitting method based on single arc segment cannot get the effective processing results.Aiming at the complexity of the operation scene,by adopting the idea of arc grouping,a wheel detection algorithm based on arc grouping and direct least square is proposed.This method firstly groups the nearest neighbor arcs locally and then determines whether the integrally grouped arcs belong to a same ellipse by setting up a threshold.Then the contours of the assembled wheel arcs are fitted by the direct least square and thus the wheels are detected by excluding the other invalid targets.Experiments show that this method can detect the wheels in complex application scenarios accurately and efficiently.Finally,the core detection algorithm,which contains the feature extraction method of truck wheels and the wheel detection algorithm,is embedded in the anti-lifting system to come up with a new anti-lifting scheme for container trucks based on video image processing.In this scheme,an abnormal decision algorithm based on wheel alteration detection is proposed and a set of anti-lifting system for container trucks is developed based on the software and hardware design.The experimental results show that this system has better timeliness and accuracy than the existing system based on laser range finder.