Performance Evaluation of Vision based Fire Detection Algorithms Using a Multi-Core Architecture

As the requirement of an efficient detection system is increasing in recent years, the need for vision based fire detection algorithms and systems is appealing. However, Vision based fire detection algorithms demand tremendous computational and I/O requirements. To support these computation algorithms, we introduce an SIMD (Single Instruction Multiple Data) based multi-core architecture that consists of 16 processing elements (PEs) and small local memories. Experimental result indicate that our multi-core architecture outperforms a high-performance TI DSP in terms of execution time(27.18x) and energy efficiency(3.89x).