An Embedded Solution for Indoor Fire Detection
The project aims to overcome the existing issues of inadequate indoor fire sensing systems (e.g., hospitals, workplaces, schools...). The goal of this project is to use security cameras to detect and recognize fires. The technique can reduce the drawbacks of a false alarm, allowing the system cost-e
2025-06-28 16:30:13 - Adil Khan
An Embedded Solution for Indoor Fire Detection
Project Area of Specialization Internet of ThingsProject SummaryThe project aims to overcome the existing issues of inadequate indoor fire sensing systems (e.g., hospitals, workplaces, schools...). The goal of this project is to use security cameras to detect and recognize fires. The technique can reduce the drawbacks of a false alarm, allowing the system cost-efficient and quick way to detect a fire, as security cameras are already deployed on the site. The framework uses Open-Source Machine Vision, or also known as OpenCV, which is an open-source freeware. An early warning feature is the major advantage of Machine Vision-based fire detection. For this framework, the camera is used for fire detection. The system processes the input feed of the camera and then operate on it for fire detection. The heat and light variations of the fire are detected in frames to identify whether a flame is involved in the process. When the fire is detected, the system sent an SMS warning to the user. It also shows the device information state on an LCD screen. Because of surveillance and insurance needs, a CCTV network comprised of cameras exists in the Indoors. Most indoors (such as hospitals, schools, workplaces, and warehouses, etc.) often normally have an extensive device for increasing security, including fire alarms and smoke detectors. In the case of a disaster, these systems may or may not be quite accurate. This system can trigger multiple false alarms because of smoke from cigarettes or incenses etc. It has enabled us to develop an algorithm that can be used to fire detection from a live video stream by using the already developed CCTV system. The expense of purchasing and maintaining complicated and ineffective fire alert system is also decreased.
Project ObjectivesThe key aim of the project is the implementation in real-time of a fire detecting technique. The key points of the project objectives are given below:
- To recognize fires when they are still small and have not grown too large.
- To build a real-time, low-cost fire detection system.
- A device that detects a fire faster and more accurately than typical fire detectors.
First, we will apply the image acquisition on the captured image from the CCTV. Using the HSI color space to transform CCTV video frames. Do segmentation of pictures. For this, the algorithm searches for an image pixel that is within 10 intensity values of the mean HSI fire value obtained from the extraction step of the feature. A value of 1 (bright pixel) is assigned to all pixels that fall within the set. (white or white) and all others are given a score of 0 (Dark pixel or black). The segmented image is used after the binary image segmentation to perform. This move renders the limits of fire evident as a picture morphological opening action (erosion accompanied by dilation). The threshold video is shown after morphological operations have been completed. The algorithm discovers contours in the segmented video stream, following morphological operations. This retrieves all the contours and reconstructs a total Nested Contour Hierarchy. It uses the horizontal compression process, Vertical segments, and diagonals, leaving just their endpoints. For each contour it has obtained, the algorithm seeks a hull. Whenever the algorithm detects some contour successfully, it sketches contour around the position of the pixel where the fire is located.
Benefits of the ProjectTo recognize fires when they are still small and have not grown too large.
To build a real-time, low-cost fire detection system.
A device that detects a fire faster and more accurately than typical fire detectors.
Technical Details of Final Deliverablethe project will be delivered as according to the requirements mentioned above.
Final Deliverable of the Project HW/SW integrated systemCore Industry EducationOther IndustriesCore Technology Internet of Things (IoT)Other TechnologiesSustainable Development Goals Quality Education, Industry, Innovation and InfrastructureRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 32510 | |||
| Raspberry Pi 4 (8 GB) | Equipment | 1 | 24000 | 24000 |
| IP Cam | Equipment | 1 | 3000 | 3000 |
| LED's | Equipment | 5 | 2 | 10 |
| LCD | Equipment | 1 | 1500 | 1500 |
| GSM | Equipment | 1 | 2500 | 2500 |
| Hard drive (120 GB) | Equipment | 1 | 1500 | 1500 |