FlameCure: An Autonomous Indoor Fire Detection and Extinguishing Surveillance Car Using Image Processing

In this paper, we project an effortless way to detect and suppress fire which is based on a machine learning algorithm that uses Open Source Computer Vision library and is also embodied by IR flame sensors and smoke detectors. Being able to use IP camera or any cloud or mobile application based image capturing device serves the security and surveillance purpose as well. The whole process offers an autonomous prototype vehicle that continues surveillance within a particular area that has a video capturing device mounted on it. Once the fire is detected, the vehicle proceeds towards it and after successfully determining and reaching the suitable stopping distance, it starts extinguishing the fire until all the fire detection mechanisms ensure complete suppression. Being able to use a variety of image capturing devices has not only paved the way of increasing its capability by using high resolution devices but also facilitated the operation for certain underdeveloped countries and low-facilitated areas. This process offers a feasible option to lessen fire based accidents at home and industries without any human involvement. Cascade classifier object detection algorithm which has been quite reliable in terms of detecting fire and distance from the prototype has been used in this project. Accuracy based on different sources of fire and camera has also been measured through real time testing. This outline can have a significant impact on automatic fire fighting that can lead to more efficient fire handling capabilities in the modern world.