An Active Early-Warning System For Bridge Piers Against Ship Collision And Its Experiment

The development of waterway transportation and the increase of bridgelead to more and more ship-bridge collision accidents. In view of thelimitation of vessel traffic service and passive system for bridge piersagainst ship collision in inland waterway, an active early-warning systemusing compound detection by visible light, infrared, laser for bridge piersagainst ship collision was proposed in this paper. The result show that theactive early-warning system can meet the need that bridge piers against shipcollision stably and reliably under all-weather, all-day, full automaticcondition.Followings are the main works and innovations in this paper：（1）A not directly measurement method was proposed which isbased on binocular vision. And the calibration algorithm andcorresponding measurement algorithm were studied in this paper. What’smore, several groups of experiments were carried out. The analysis of themeasurement error shows that the algorithms can meet the demands ofsystem application. The algorithms solved the problem that it isinconvenient to directly measure the ship in inland waterway.（2）Detection algorithms were studied in this paper. Since there aredetection noise of virtual object, background subtraction algorithm wasimproved. What’s more, experiments and performance analysis werecarried out. The result shows that improved algorithm can effectively solvethis problem. The camshift algorithm which is based on the colorinformation was used for target detection and tracking in the daylight.Since the information of visible light is not enough at night or in rainy and greasy weather, the IR image is used for object detection and tracking. Thecamshift algorithm which is based on the texture information was used forobject tracking because of the feature of IR image. The result shows thatthe system can warn the ships that will strike the bridge piers withacousto-optic stably and reliably under all-weather, all-day, full automaticcondition. So it will decrease ship-bridge collision accidents and protectthe bridge and ship. Kalman algorithm was used for movement estimationin this paper. We use the value which is predicted by kalman as theinitialization of camshift. And the experiment result showed that thismethod can reduce the iterations.（3）The prototype of the active early warning system for bridge piersagainst ship collision has been built based on photoelectricity detection inthis paper. The software system has been designed and accomplished. Andthe experimental study and performance analysis show that it will be morereliable for bridge piers against ship collision by using combination of theactive system and passive system.