Algorithm of fire detection for multi-sensor system

Abstract

Objective. The paper proposes a fire detection algorithm for a multisensor system. Due to the difficult conditions in the field, for the first time, watch rescue operations are difficult and often endanger the lives of rescuers. The main scientific goal is for the system to work autonomously on certain segments of the monitoring process, while the three agents, to varying degrees, must interact with each other based on communication and decision-making algorithms.Method. There are many algorithms for image processing, but algorithms that use several sources of information are not sufficiently developed and described. The focus is on fire detection algorithms. The algorithm was developed using NI Vision Assistant, a software tool for rapid prototyping and testing of imaging applications.Result. In addition to the software implementation in C, which NI Vision Assistant generates by default, the paper presents a Python implementation of the algorithm.Conclusion. The results of the work can be used to develop multisensor systems for monitoring hard-to-reach areas.

Keywords

• image processing
• fire detection
• image segmentation
• multi-agent multisensor system