Back to Search
Start Over
A Novel Efficient Video Smoke Detection Algorithm Using Co-occurrence of Local Binary Pattern Variants.
- Source :
- Fire Technology; Sep2022, Vol. 58 Issue 5, p3139-3165, 27p
- Publication Year :
- 2022
-
Abstract
- Smoke detection is an advance caution to the unforeseen great damage events. Therefore, it is required to identify the smoke in the course of initial stages for preventing fire events. A new technique is proposed to lessen the rate of incorrect alarm by identify the smoke and examine its distinctive texture attributes. Initially, the smoke-colored regions are segmented based on color at the YUV color locality. Then the tentative frame differencing is used to segment the candidate smoke region from the smoke-colored region. In the next phase, the candidate distinctive texture attributes in the smoke region are extracted using Co-occurrence of Hamming Distance based Local Binary pattern (CoHDLBP) and Co-occurrence of Local Binary pattern (CoLBP); these features include homogeneity, energy, correlation and contrast. Finally, the ELM classifier is proficient for the take-out features from the candidate smoke region, and then the decision has been taken with the assistance of a smoke alarm. Investigational outcomes proved that the suggested smoke recognition process executes better compared with all the usual smoke recognition methods by achieving better detection accuracy and processing time. [ABSTRACT FROM AUTHOR]
- Subjects :
- FIRE detectors
ALARMS
SMOKE
HAMMING distance
ALGORITHMS
Subjects
Details
- Language :
- English
- ISSN :
- 00152684
- Volume :
- 58
- Issue :
- 5
- Database :
- Complementary Index
- Journal :
- Fire Technology
- Publication Type :
- Academic Journal
- Accession number :
- 159213073
- Full Text :
- https://doi.org/10.1007/s10694-022-01306-2