Research on spatial distribution of ship encounters based on optimized feature extraction algorithm and integrated AIS data

The research on the spatial distribution relationship of ship encounters holds significant importance for ensuring navigation safety. Against this backdrop, heightened demands are placed on the situation awareness between ships. This paper, focusing on the complex and ever-changing navigation environment at sea, proposes an improved LBP (Local Binary Pattern) algorithm for feature extraction of target ships. This algorithm first establishes a background model based on the mode of grayscale values in the histogram of each pixel across image sequences, enabling it to adapt to scene changes. Subsequently, an enhanced NRLBP-U (Non-Redundant Local Binary Pattern-Uniform) algorithm is employed to extract ship features. Then, the inter-frame difference method is utilized for real-time target tracking of other ships. Finally, AIS (Automatic Identification System) data is integrated to construct the spatial distribution relationship of ship encounters. Experimental results demonstrate that this algorithm can effectively extract ship features in maritime moving target detection with robust performance. Furthermore, the optimized extraction algorithm, when fused with AIS data, can more accurately assess the navigation situation of target ships and establish a more precise spatial distribution relationship of ship encounters.