构建基于智能视觉的船舶碰撞风险评估模型,以提升船舶航行安全性。利用机器视觉技术采集的船舶航行图像,提取船舶航行图像中的颜色特征与梯度方向特征。利用高效卷积运算符(efficient convolution operators for tracking,ECO)算法学习...构建基于智能视觉的船舶碰撞风险评估模型,以提升船舶航行安全性。利用机器视觉技术采集的船舶航行图像,提取船舶航行图像中的颜色特征与梯度方向特征。利用高效卷积运算符(efficient convolution operators for tracking,ECO)算法学习滤波器模板,获取船舶航行图像特征的响应图值。加权求和处理特征响应图值,实现船舶航行轨迹跟踪。依据船舶航行轨迹跟踪结果的时间与坐标信息,获取船舶航行的速度、方位以及坐标信息。依据船舶航行速度等信息,构建船舶碰撞风险评估模型,输出船舶碰撞风险值。测试结果表明,该模型有效评估船舶航行过程中的对遇碰撞、交叉碰撞以及追越碰撞风险,提升船舶航行安全性。展开更多
Despite of modern navigation devices, there are problems in navigation of vessels in waterways due to the geographical structures, disturbances in water, dynamic nature, and heavily environmental influenced sea traffi...Despite of modern navigation devices, there are problems in navigation of vessels in waterways due to the geographical structures, disturbances in water, dynamic nature, and heavily environmental influenced sea traffic. Even though all vessels are equipped with modern navigation devices, the accidents are reported caused by various reasons and mainly by human factor according to investigation. We propose an effective and efficient composition collision risk calculation method for finding the collision probability and avoiding the collision between ships in possible collision situations. The proposed composition collision risk calculation method at ship's position using combination of fuzzy and fuzzy comprehensive evaluation methods. The algorithm is straightforward to implement and is shown to be effective in automatic ship handling for ships involved in complex navigation situations. Experiments are carried out with indigenous data and the results show the effectiveness of the proposed approach.展开更多
文摘构建基于智能视觉的船舶碰撞风险评估模型,以提升船舶航行安全性。利用机器视觉技术采集的船舶航行图像,提取船舶航行图像中的颜色特征与梯度方向特征。利用高效卷积运算符(efficient convolution operators for tracking,ECO)算法学习滤波器模板,获取船舶航行图像特征的响应图值。加权求和处理特征响应图值,实现船舶航行轨迹跟踪。依据船舶航行轨迹跟踪结果的时间与坐标信息,获取船舶航行的速度、方位以及坐标信息。依据船舶航行速度等信息,构建船舶碰撞风险评估模型,输出船舶碰撞风险值。测试结果表明,该模型有效评估船舶航行过程中的对遇碰撞、交叉碰撞以及追越碰撞风险,提升船舶航行安全性。
基金supported by ETRI through Maritime Safety & Maritime Traffic Management R&D Program of the MOF/KIMST (2009403, Development of Next Generation VTS for Maritime Safety)supported by the National Research Foundation of Korea (NRF) Grant funded by the Korea government (MEST) (No. 2011-0015009)
文摘Despite of modern navigation devices, there are problems in navigation of vessels in waterways due to the geographical structures, disturbances in water, dynamic nature, and heavily environmental influenced sea traffic. Even though all vessels are equipped with modern navigation devices, the accidents are reported caused by various reasons and mainly by human factor according to investigation. We propose an effective and efficient composition collision risk calculation method for finding the collision probability and avoiding the collision between ships in possible collision situations. The proposed composition collision risk calculation method at ship's position using combination of fuzzy and fuzzy comprehensive evaluation methods. The algorithm is straightforward to implement and is shown to be effective in automatic ship handling for ships involved in complex navigation situations. Experiments are carried out with indigenous data and the results show the effectiveness of the proposed approach.
