In this paper, we address the problem of forest fires’ early detection and segmentation in order to predict their spread and help with fire fighting. Techniques based on Convolutional Networks are the most used and have proven to be efficient at solving such a problem. However, they remain limited in modeling the long-range relationship between objects in the image, due to the intrinsic locality of convolution operators. In order to overcome this drawback, Transformers, designed for sequence-to-sequence prediction, have emerged as alternative architectures. They have recently been used to determine the global dependencies between input and output sequences using the self-attention mechanism. In this context, we present in this work the very first study, which explores the potential of vision Transformers in the context of forest fire segmentation. Two vision-based Transformers are used, TransUNet and MedT. Thus, we design two frameworks based on the former image Transformers adapted to our complex, non-structured environment, which we evaluate using varying backbones and we optimize for forest fires’ segmentation. Extensive evaluations of both frameworks revealed a performance superior to current methods. The proposed approaches achieved a state-of-the-art performance with an F1-score of 97.7% for TransUNet architecture and 96.0% for MedT architecture. The analysis of the results showed that these models reduce fire pixels mis-classifications thanks to the extraction of both global and local features, which provide finer detection of the fire’s shape.

In this paper, we present a new enhancement for an emergency and disaster relief system called Critical and Rescue Operations using Wearable Wireless sensors networks (CROW ${}^2$ ). We address the reliability challenges in setting up a wireless autonomous communication system in order to offload data from the disaster area (rescuers, trapped victims, civilians, media, etc.) back to a command center. The proposed system connects deployed rescuers to extended networks and the Internet. CROW ${}^2$ is an end-to-end system that runs the recently-proposed Optimized Routing Approach for Critical and Emergency Networks (ORACE-Net) routing protocol. The system integrates heterogeneous wireless devices (Raspberry Pi, smart phones, sensors) and various communicating technologies (WiFi IEEE 802.11n, Bluetooth IEEE 802.15.1) to enable end-to-end network connectivity, which is monitored by a cloud Internet-of-Things platform. First, we present the CROW ${}^2$ generic system architecture, which is adaptable to various technologies integration at different levels (i.e., on-body, body-to-body, off-body). Second, we implement the ORACE-Net protocol on heterogeneous devices including Android-based smart phones and Linux-based Raspberry Pi devices. These devices act as on-body coordinators to collect information from on-body sensors. The collected data is then pushed to the command center thanks to multi-hop device-to-device communication. Third, the overall CROW ${}^2$ system performance is evaluated according to relevant metrics including end-to-end link quality estimation, throughput and end-to-end delay. As a proof-of-concept, we validate the system architecture through deployment and extracted experimental results. Finally, we highlight motion detection and links’ unavailability prevention based on the recorded data where the main factors (i.e., interference and noise) that affect the performance are analyzed.

With the widespread use of deep learning in leading systems, it has become the mainstream in the table detection field. Some tables are difficult to detect because of the likely figure layout or the small size. As a solution to the underlined problem, we propose a novel method, called DCTable, to improve Faster R-CNN for table detection. DCTable came up to extract more discriminative features using a backbone with dilated convolutions in order to improve the quality of region proposals. Another main contribution of this paper is the anchors optimization using the Intersection over Union (IoU)-balanced loss to train the RPN and reduce the false positive rate. This is followed by a RoI Align layer, instead of the ROI pooling, to improve the accuracy during mapping table proposal candidates by eliminating the coarse misalignment and introducing the bilinear interpolation in mapping region proposal candidates. Training and testing on a public dataset showed the effectiveness of the algorithm and a considerable improvement of the F1-score on ICDAR 2017-Pod, ICDAR-2019, Marmot and RVL CDIP datasets.