ALGORITHM FOR DETECTING FIRES IN AGRICULTURAL FIELDS

Lyakhmanov Dmitry Alexandrovich, Kapranov Sergey Nikolaevich, Stepanenko Maria Anatolievna

This article discusses search algorithms for smoke-like structures in a video stream in order to identify fires in open agricultural areas. The principle of the search is based on the temporal analysis of the change in the brightness index of the pixel groups of the video stream. For pixel groups belonging to the video stream of combustion products, an irregular smooth change in brightness is characteristic, which makes it possible to distinguish them from other dynamic objects, such as clouds or moving objects. The proposed algorithm makes it possible to determine smoke-like structures from a video stream, based on video streams of extremely low resolutions. This advantage makes it possible to detect fires at large distances, without the use of high-resolution video cameras. The test results showed the consistency of the proposed algorithm and the possibility of its use in video surveillance systems for the state of agricultural areas.

Keywords:Fire detection, agricultural areas, video stream analysis, smoke-like structures, low resolution