作者: Zhongli Ma , Jie Wen , Xiumei Liang
DOI: 10.1007/978-3-642-38715-9_52
关键词: Tracking (particle physics) 、 Marine layer 、 Computer vision 、 Diffuse sky radiation 、 Sky 、 Algorithm 、 Artificial intelligence 、 Image segmentation 、 Accuracy and precision 、 Brightness 、 Computer science 、 Identification (information)
摘要: The visual system is one of the main equipment unmanned surface vehicle (USV) autonomous navigation. Under sea fog, atmospheric particles scattering leads to serious image degradation system. Because there obvious sea-sky-line and larger sky area in offshore, so firstly, segmentation done get area, through anglicizing characteristics, brightness estimated, then a simplified physical model built up, lastly scene recovery finished. Thinking about using this simple defogging method video image, foreground background separation done. Comparative research with several methods onshore, results show that proposed can enhance clarity USV under fog very well. This brought good foundation further improve accuracy precision target identification tracking algorithm.