Ship license plate detection method in natural scene

Abstract

The invention provides a ship license plate detection method in a natural scene. The method comprises the following steps: 1) data acquisition: collecting videos of ships entering and exiting a port through a camera; 2) data processing: processing the video, and screening out a picture containing a ship license plate; 3) data annotation: carrying out ship license plate annotation on the picture; 4) model training: using a target detection model to train ship license plate detection; and 5) model detection: detecting the picture by using the trained ship license plate detection model to obtainthe position of each ship license plate in the picture. A deep learning model R2CNN is selected, the multi-angle ship license plate can be identified, the detection speed and accuracy of the ship license plate are improved, and the method has high practical application value.

Description

technical field [0001] The invention relates to the field of computer application technology and the field of target detection, in particular to a ship plate detection method in a natural scene. Background technique [0002] In recent years, with the increasingly prosperous development of the port economy, the scale of fishery engineering and shipping engineering has continued to expand, and the number of ships has continued to increase. At the same time, illegal smuggling and illegal fishing at sea have also emerged. Since there are often situations where AIS is manually turned off to evade supervision, traditional manual detection is time-sensitive and dangerous, making it difficult to achieve effective supervision of ships. In order to strengthen the dimension and strength of management, a large number of high-definition surveillance cameras have been set up in the port area, which can capture video images of ships at sea from a long distance and provide direct evidence f...

AI Technical Summary

Problems solved by technology

Due to the lack of unified ship plate standards, the hanging position of the ship plate is not uniform, the background color and character color of the ship plate are varied, and the image of the ship plate captured by the camera is ideally a rectangle. However, when the ship passes the channel, its position is different The angle change between the camera and the ship plate caused by uncertainty and the placement of the camera device will inevitably cause obvious distortion and tilt of the captured image

Such different degrees of image deformation will have a negative impact on the detection system, which will easily cause mis-segmentation of characters, resulting in a decline in the recognition rate of ship plates

Therefore, it is difficult to directly transplant license plate detection technology to ship plate detection, resulting in less research on ship plate detection and immature technology

[0004] Therefore, in order to solve the above problems, a ship plate detection method based on deep learning is needed, which can avoid the technical problems of low precision, slow speed and poor effect in the ship plate detection technology in the prior art

Images