Video analysis of the degassing process

Abstract

This thesis presented a collaborative effort between Ovako and Swerim to create a system to assist operators decision-making during the degassing stage of Ovako’s steel production, using video data obtained from the degassing camera feed. The purpose of the project was to explore all potential information that could be extracted from the camera feed. This would act as proof-of-concept for such a system. The research goal were accomplished by employing image analysis techniques and statistical methods to collect as much relevant information as possible. Traditional image analysis techniques were utilized for segmentation and feature extraction, which were combined with statistical methods to estimate key degassing properties such as fluid height/volume, fluid movement, and proportions of steel/slag in the degassing ladle. The results demonstrated a functional system which processed saved video data, providing estimated statistics at a chosen frame rate and producing relevant plots showing the segmentation and statistics. Findings demonstrated ample opportunities for extracting valuable information from the degassing images that exceeded initial expectations set forth by engineers at Ovako and Swerim. Additionally, a correlation was seen between extracted statistics from the system and actual video footage, suggesting the reliability of the system. This thesis demonstrated the potential of creating a comprehensive system to aid the operators during degassing. It served as a starting point for further discussions among the companies regarding refinements to align more closely with operator needs. More work must be done towards realising the goal, yet this thesis provided proof of concept that the development of such a system is feasible.