Research on ship speed prediction based on time series imaging and deep convolutional network fusion method

Ship speed prediction is crucial for ship operation and safe navigation, providing precise data support for route planning and optimization, port operational efficiency enhancement, navigation system development, etc. Currently, the computational accuracy of the existing ship speed prediction methods faces great challenges. To enhance the precision and efficiency of ship speed prediction, an innovative method applying time series imaging technology for ship speed prediction is proposed. In this method, time series data is first sliced and converted into two-dimensional images. The data points of each period are converted into pixels in the image, and the intensity or color depth of the pixels represents the corresponding size of the data values. Subsequently, a deep convolutional network prediction model is constructed and trained to predict ship speed accurately. Finally, the effectiveness of the proposed method is validated using two sets of ship speed data, with a mean square error of 0.08 on the open-source dataset and 0.026 on the real ship dataset. The results indicate that the proposed method can achieve accurate prediction of ship speed.