Currently, sensor-based systems for fire detection are widely used worldwide. Further research has shown that camera-based fire detection systems achieve much better results than sensor-based methods. In this study, we present a method for real-time high-speed fire detection using deep learning. A new special convolutional neural network was developed to detect fire regions using the existing YOLOv3 algorithm. Due to the fact that our real-time fire detector cameras were built on a Banana Pi M3 board, we adapted the YOLOv3 network to the board level. Firstly, we tested the latest versions of YOLO algorithms to select the appropriate algorithm and used it in our study for fire detection. The default versions of the YOLO approach have very low accuracy after training and testing in fire detection cases. We selected the YOLOv3 network to improve and use it for the successful detection and warning of fire disasters. By modifying the algorithm, we recorded the results of a rapid and high-precision detection of fire, during both day and night, irrespective of the shape and size. Another advantage is that the algorithm is capable of detecting fires that are 1 m long and 0.3 m wide at a distance of 50 m. Experimental results showed that the proposed method successfully detected fire candidate areas and achieved a seamless classification performance compared to other conventional fire detection frameworks.

Wildfire is one of the most significant dangers and the most serious natural catastrophe, endangering forest resources, animal life, and the human economy. Recent years have witnessed a rise in wildfire incidents. The two main factors are persistent human interference with the natural environment and global warming. Early detection of fire ignition from initial smoke can help firefighters react to such blazes before they become difficult to handle. Previous deep-learning approaches for wildfire smoke detection have been hampered by small or untrustworthy datasets, making it challenging to extrapolate the performances to real-world scenarios. In this study, we propose an early wildfire smoke detection system using unmanned aerial vehicle (UAV) images based on an improved YOLOv5. First, we curated a 6000-wildfire image dataset using existing UAV images. Second, we optimized the anchor box clustering using the K-mean++ technique to reduce classification errors. Then, we improved the network’s backbone using a spatial pyramid pooling fast-plus layer to concentrate small-sized wildfire smoke regions. Third, a bidirectional feature pyramid network was applied to obtain a more accessible and faster multi-scale feature fusion. Finally, network pruning and transfer learning approaches were implemented to refine the network architecture and detection speed, and correctly identify small-scale wildfire smoke areas. The experimental results proved that the proposed method achieved an average precision of 73.6% and outperformed other one- and two-stage object detectors on a custom image dataset.

The technologies underlying fire and smoke detection systems play a crucial role in ensuring and delivering optimal performance in modern surveillance environments. In fact, fire can cause significant damage to lives and properties. Considering that the majority of cities have already installed camera-monitoring systems, this encouraged us to take advantage of the availability of these systems to develop cost-effective vision detection methods. However, this is a complex vision detection task from the perspective of deformations, unusual camera angles and viewpoints, and seasonal changes. To overcome these limitations, we propose a new method based on a deep learning approach, which uses a convolutional neural network that employs dilated convolutions. We evaluated our method by training and testing it on our custom-built dataset, which consists of images of fire and smoke that we collected from the internet and labeled manually. The performance of our method was compared with that of methods based on well-known state-of-the-art architectures. Our experimental results indicate that the classification performance and complexity of our method are superior. In addition, our method is designed to be well generalized for unseen data, which offers effective generalization and reduces the number of false alarms.

Automatic extraction of salient regions is beneficial for various computer vision applications, such as image segmentation and object recognition. The salient visual information across images is very useful and plays a significant role for the visually impaired in identifying tactile information. In this paper, we introduce a novel saliency cuts method using local adaptive thresholding to obtain four regions from a given saliency map. First, we produced four regions for image segmentation using a saliency map as an input image and local adaptive thresholding. Second, the four regions were used to initialize an iterative version of the GrabCuts algorithm and to produce a robust and high-quality binary mask with a full resolution. Finally, salient objects’ outer boundaries and inner edges were detected using the solution from our previous research. Experimental results showed that local adaptive thresholding using integral images can produce a more robust binary mask compared to the results from previous works that make use of global thresholding techniques for salient object segmentation. The proposed method can extract salient objects with a low-quality saliency map, achieving a promising performance compared to existing methods. The proposed method has advantages in extracting salient objects and generating simple, important edges from natural scene images efficiently for delivering visually salient information to the visually impaired.

The growing aging population suffers from high levels of vision and cognitive impairment, often resulting in a loss of independence. Such individuals must perform crucial everyday tasks such as cooking and heating with systems and devices designed for visually unimpaired individuals, which do not take into account the needs of persons with visual and cognitive impairment. Thus, the visually impaired persons using them run risks related to smoke and fire. In this paper, we propose a vision-based fire detection and notification system using smart glasses and deep learning models for blind and visually impaired (BVI) people. The system enables early detection of fires in indoor environments. To perform real-time fire detection and notification, the proposed system uses image brightness and a new convolutional neural network employing an improved YOLOv4 model with a convolutional block attention module. The h-swish activation function is used to reduce the running time and increase the robustness of YOLOv4. We adapt our previously developed smart glasses system to capture images and inform BVI people about fires and other surrounding objects through auditory messages. We create a large fire image dataset with indoor fire scenes to accurately detect fires. Furthermore, we develop an object mapping approach to provide BVI people with complete information about surrounding objects and to differentiate between hazardous and nonhazardous fires. The proposed system shows an improvement over other well-known approaches in all fire detection metrics such as precision, recall, and average precision.