【摘要】 针对船舶监控中AIS与视频图像跟踪轨迹关联匹配难的问题，提出一种基于综合因素模糊评判算法的AIS与视频图像融合方法.设计了分布式框架融合模型，通过Darknet网络模型结合YOLOv3算法深度学习框架，实现视频图像中船舶目标的识别与跟踪.抽取船舶AIS的目标运动特征，通过短时时空轨迹预测和坐标转换实现AIS航迹与视频图像跟踪轨迹的时空配准.基于综合因素模糊评判算法实现船舶AIS与监控图像信息融合.结果表明：所提出的方法对内河航道多船舶目标的融合准确率达到82.4%,对船舶交汇场景依据具有很高的识别率.更多还原

【Abstract】 Aiming at the problem that it is difficult to match the tracking trajectory of AIS and video images in ship monitoring, a fusion method of AIS and video images based on comprehensive factor fuzzy evaluation algorithm was proposed. A distributed framework fusion model was designed, and the identification and tracking of ship targets in video images were realized by combining Darknet network model with YOLOv3 algorithm deep learning framework. The target motion characteristics of ship AIS were extracted, and the temporal and spatial registration of AIS track and video image tracking track was realized through short-term temporal and spatial trajectory prediction and coordinate transformation. The fusion of ship AIS and monitoring image information was realized by fuzzy evaluation algorithm based on comprehensive factors. The results show that the fusion accuracy of the proposed method for multi-ship targets in inland waterways reaches 82.4%, and it has a high recognition rate for the basis of ship intersection scenes.更多还原