z-logo
open-access-imgOpen Access
Machine learning technologies for automated draft measurements
Author(s) -
A.N. Ivanovsky,
AUTHOR_ID,
Ye.G. Zinchenko,
S.G. Cherny,
AUTHOR_ID,
AUTHOR_ID
Publication year - 2021
Publication title -
trudy krylovskogo gosudarstvennogo naučnogo centra
Language(s) - English
Resource type - Journals
eISSN - 2618-8244
pISSN - 2542-2324
DOI - 10.24937/2542-2324-2021-2-s-i-33-39
Subject(s) - hull , inclinometer , computer science , artificial intelligence , process (computing) , work (physics) , machine learning , industrial engineering , engineering , marine engineering , mechanical engineering , cartography , geography , operating system
This paper discusses the process of ship draft measurement known as draft survey. The purpose of the study was to improve the accuracy of draft survey results and the efficiency of this procedure itself. The study relies on video footages of draft marks, as well as clinometer readings, following the methods of digital image processing, machine learning, digital signal processing, linear filtering and applied programming. The tools developed as a result of this work are based on machine-learning algorithms and can perform draft surveys even in bad weather. Accuracy limits depend on camera resolution, lighting and weather conditions. Combined with linear filtering algorithms and ship inclinometers, this technology might offer draft survey tolerances as narrow as several millimeters, thus being well above its existing counterparts. Automated draft survey method suggested in this paper will make cargo weight measurements of bulkers more accurate, thus saving time and money, as well as making survey results independent on human error. Relying on machine-learning and computer-vision technologies, this method is universal and will work with any type of ships. Theoretical value of this study is that it gives a comprehensive review of what ship draft is and how it is measured.

The content you want is available to Zendy users.

Already have an account? Click here to sign in.
Having issues? You can contact us here