Spatiotemporal bag-of-features for early wildfire smoke detection

作者：

Highlights：

• We select key frames from a video and detect candidate blocks only in key frames.

• We prepare 3D spatiotemporal volumes by combining the candidate blocks.

• We introduce a new weighting scheme for generating a more reasonable BoF.

• The random forest classifier is built during the training phase by using the BoF.

摘要

•We select key frames from a video and detect candidate blocks only in key frames.•We prepare 3D spatiotemporal volumes by combining the candidate blocks.•We introduce a new weighting scheme for generating a more reasonable BoF.•The random forest classifier is built during the training phase by using the BoF.

论文关键词：Wildfire smoke detection,Spatiotemporal feature,Bag-of-features,Histogram of oriented gradient,Histogram of oriented optical flow,Random forest

论文评审过程：Received 25 January 2013, Revised 24 June 2013, Accepted 12 August 2013, Available online 19 August 2013.

论文官网地址：https://doi.org/10.1016/j.imavis.2013.08.001