A Two-stage Method for Ship Target Detection and Identification in Complex Marine Environment

Ships are the main means of transportation at sea,and the hull number is the “identity card” of the ship. Therefore, it is necessary to effectively supervise ships through the identification of hull numbers. Due to the complex environment of the ocean and shooting angle limitations, hull number images are usually blurred and distorted, which has a great impact on the recognition results. Therefore, we propose a two-stage method for the detection and recognition of hull numbers, and a Super Resolution-Correction Recognition network is proposed for ambiguity and distortion of hull numbers. The accuracy of hull number detection reaches 92.1%, and the accuracy of hull number recognition reaches 91.3%. Experiments show that the method in this paper can effectively identify hull numbers.