Saliency-based Video Flame Recognition Method for Construction Site

摘要

With the rapid development of economic construction, the number of construction sites increases significantly, and fires and losses have shown an upward tendency year by year. In order to prevent and monitor fire accidents, saliency-based video flame recognition method for construction sites is proposed. First of all, the suspected flame areas are extracted using the visual saliency detection algorithm. Then, the suspected region is segmented in the YCbCr color space. When a suspected flame segmentation area is identified, the change of flame area in the segmentation area is analyzed to finally determine whether there is a fire accident.