Mitsuru KOBAYASHIKeiji SATO
28-29页查看更多>>摘要：We developed a system that detects other ships and estimates their relative positions from images simultaneously captured by two visible-light cameras mounted on a commercial ship facing forward. Deep learning was used for ship detection, and stereo vision methods were employed for position estimation. The system detected 95% of ship images with a size of 32 pixels or more, and the position estimation error is comparable to the theoretically predicted value.
原文链接:
NETL
NSTL
