摘要
为克服当前驾驶台值班报警系统需人工直接参与的不足,提出一种基于视频流跟踪的驾驶台值班报警系统。该系统通过比对2时刻值班人员的位置差异来判定值班人员的清醒状态。使用定时图像差分的方法确定值班人员的图上位置,实现目标的图上坐标与驾驶台空间坐标的转换;系统所收集数据可用于后续值班行为分析使用。实船实验表明,系统克服了需人员直接参与的限制,能够根据所设定的时间间隔识别驾驶员值班状态,并据此发出系列报警信号。
A new design of bridge navigational watch alarm system （ BNWAS） was introduced to overcome the drawbacks of current systems, which required the engagement of watch-keepers on bridges of ships. The watch-keepers ’ awake state was judged by comparing the difference of two locations at different time. The image differencing was used as an approach to locate the target and the target,s location on the image was further transferred to its space coordinates in the bridge. Meanwhile, the collect-ed data could be used for the analyzing of human element on the maritime accidents. Experiment on an operating ship showed that the limitation of watch-keepers ’ engagement could be eliminated significantly and the watching states of navigators could be moni-tored and assessed.
出处
《船海工程》
北大核心
2017年第2期170-174,共5页
Ship & Ocean Engineering
基金
国家自然科学基金(51279099)
上海市科学技术委员会基金(12ZR1412500)
上海市教委科研创新基金资助重点项目(13ZZ124)
上海市教育委员会和上海市教育发展基金会"曙光计划"基金(12SG40)
交通运输部应用基础研究(2013329810300)
