An Embedded Cellular Automata Algorithm for Infrared Object Detection

This work proposes and presents a cellular automata algorithm for infrared object detection and tracking. The algorithm’s application is aimed towards computationally limited edge computing sensory modules. Performed is a statistical evaluation of the algorithm’s noise resilience through a 3D histogram method. Proposed is a methodology for the optimal choice of the algorithm’s fine-tuning parameter values. The performance influence of applying low-pass Kernel filtering was evaluated and developed was a majority voting-based concurrent object mask-generating algorithm.