The detection of steam injection based on video surveillance

In this paper, we focus on the visual features of steam injection and propose an integrated algorithm to detect it based on video surveillance. The proposed method is depended on three decision rules which are the attribute of gray level and the feature of frequent flicker rate of steam injection, and the similarity structure between background image and current frame. The block-based approach is applied to all three decision rules. The experimental results show that the algorithm provides a reliable detection method which is useful in many cases such as the alarm on the leakage of a heating pipe.