Research and application of ship identification based on image processing

Ship name identification carries identity information. Ship data can be collected quickly by identifying the identification characters in the ship image. However, due to many factors such as light pollution, large tilt angle and environmental interference in ship image, ordinary Optical Character Recognition is difficult to process directly. Based on the self-built ship image dataset, this paper completes the preprocessing of the target image by means of gray levelization, straight square equalization and noise removal. Afterwards, the improved character region positioning algorithm and the fusion neural network model are used to realize the extraction of ship number characters. The recognition accuracy of ship number character can reach 92%. At the same time, referring to the method of highway traffic management, this identification method is used as the core function to design the ship identification system.