Abstract
Recently, the avoidance from ship collisions becomes a mandatory problem in the viewpoint of maneuvering ships safety in crowded or crossing areas. Today, information for the steering is mainly limited to pilot's eye measurement and radar information.
As for radar information, the detecting the course direction or turning of a target vessel urgently and accurately with an accuracy is difficult. Moreover, information on the bigness and the altering course of a target vessel obtained by eye measurement is regarded as seriously important for the avoidance. Then, the authors have studied on the collision avoidance systems that make use of image information instead of eye measurement.
In this study, they obtain a method for measuring course of a target vessel and evaluation of the collision risk by the extracted image of target from IR image.
As the result of this study;
(1) The image of the vessel is extracted from IR image base on the threshold set method by using area information.
(2) The course and the length of a target vessel are obtained from the number of the pixels corresponding to horizontal length of the vessel. And the altering course of a target vessel is detected on real time.
(3) As for the direction of the prow of the target vessel, it is done from the specified centroid of the vessel image.
(4) The amount of the altering course necessary for avoidance is obtained from the meeting condition predicted from radar and the image. And effect is evaluated.
It is demonstrated and evaluated that the effective and simultaneous use of radar and the image can improve the performance of collision avoidance.
