Ship icon image superposition error correction method

Abstract

The invention discloses a ship icon image superposition error correction method, which comprises the steps of 1, acquiring multiple frames of icon images of a target ship at the same moment, and performing noise reduction processing to obtain deblurred images; 2, superposing the deblurred images, and binarizing the deblurred images to output image data; 3, clustering the image data by using a neural network method and dividing the image data into a plurality of class region data; and 4, removing discrete noise of a class region by adopting two-dimensional Gaussian filtering to obtain corrected images, and splicing the corrected images to obtain a complete ship icon image. According to the method, superposition errors of the ship icon image are corrected through the processes of noise reduction and deblurring, image correction and superposition and cluster noise reduction to remove discrete noise points of the overlapped part, and the obtained ship icon image is high in resolution and precision.

Description

technical field [0001] The invention relates to the technical field of ship icon image recognition, in particular to a method for correcting superposition errors of ship icon images. Background technique [0002] In military and civilian fields such as maritime police, maritime affairs, and ship management systems, image monitoring has become an important source of information. It can quickly and effectively extract ship targets from real-time monitoring, automatically identify ship identities, and alert ships to abnormal behavior. become an important field of application. [0003] In the target detection technology, the image fusion technology after multi-angle shooting is often adopted to recognize the icon image. Due to the influence of the environment, measurement error, and different fields of view, there are differences in the images obtained from different angles, and the multiple images are directly superimposed on the same image. After a certain area, there will be...

AI Technical Summary

Problems solved by technology

[0003] In target detection technology, image fusion technology after multi-angle shooting is often adopted to identify icon images. Due to the influence of environment, measurement errors, and different fields of view, there are differences in images acquired from different angles, so multiple images are directly superimposed on the same image. After a certain area, there will be large errors in the image, and the image quality, resolution and sensitivity will be affected

Images