Thesis Defended (3/4/19) – Waiting for Graduation

PS: Sorry for the low quality photo, we just downloaded the photo via Facebook Messenger

Smart Fire Detection and Sprinkler System

Abstract

Most of the main problem(s) encountered by the firefighters and the authorities is the lack of message automation, and the delay of activation of smoke sensors which would cause more damage if there is no action taken from the authorities. In this problem, the team built Smart Fire Detector and Sprinkler System that detects fire/smoke by the use of camera, it automatically turns on the buzzer and the Sprinkler system if there is a fire detected and to be followed with a warning message that contains the details and information of the user’s name, the address and its coordinates which is the location.

The researches uses two (2) raspberry pi model B and B+, a tiny credit card sized and affordable computer that can be plugged in any TV/monitor, (2) logictech c270 usb web cameras, an affordable quality camera with a 720p/30fps video quality and a 60 degree field of view, an Arduino nano, a small, complete and breadboard-friendly board to connect and .communicate with other modules such as, the gsm module, the buzzer for the fire alarm system, and the pump which is also connected to a relay from the Arduino nano. During the testing, first we trained and captured a video file, snapshot and cropped hundreds of images that has a similar characteristics of fire and a smoke using the Logitech usb-web camera. We use the algorithm called deep learning haar-cascading algorithm, is an object detection algorithm used to locate faces, pedestrians, objects and facial expressions in an image. After training the images, the researchers will begin another test(s) of five (5) or more to test the system’s accuracy of detecting fire/smoke.

The results are likely to detect fire ranging no(s). (1-5) is 4 that is 80% accurate from the current five (5) testing’s, and somewhat likely to detect smoke ranging no(s). (1-5) which is 2 that is 20% accurate from the current five (5) fire and smoke testing’s. Therefore, Smart Fire Detector and Sprinkler System is effective and accurate in terms of detecting fire. In addition, the authorities will also have an automated warning message.

Leave a Reply

Your email address will not be published. Required fields are marked *