Monitoring of oil leakage from a ship propulsion system using IR camera and wavelet analysis for prevention of health and ecology risks and engine faults

It is a well known fact that oil leakage from ship diesel engines is harmful both for the environment and the ship engine and therefore has to be observed and alarmed. The present paper proposes a system for overcoming described problems by installing a computer vision system. The used algorithm of pattern recognition system is based on the use of wavelet structures. Additionally, one of the problems in the system is the compensation of camera movements due to engine vibration. The compensation part of the computer vision solution is used to improve position determination. The position determination is improved more that 300 % when using farras wavelets.