Smoke Detection Using Temporal Hoghof Descriptors and Energy Colour Statistics from Video

We present a novel algorithm for detecting and localizing smoke in video. The first step of our method focuses on detecting the presence of smoke in video frames, while in the second part localization of smoke particles in the scene takes place. In our implementation, we take advantage of both appearance and motion information, so that we can extract robust and meaningful information. Machine learning is used in order to discriminate our data more thoroughly and provide accurate smoke detection. Experiments carried out with various benchmark datasets show that smoke is indeed accurately localized both in time and space via the proposed approach.