Terrain-influenced incremental watchtower expansion for wildfire detection

A Forest Fire Recognition Method Using UAV Images Based on Transfer Learning

Timely detection of forest wildfires is of great significance to the early prevention and control of large-scale forest fires. Unmanned Aerial Vehicle(UAV) with cameras has the characteristics of wide monitoring range and strong flexibility, making it very suitable for early detection of forest fire. However, the visual angle/distance of UAV in the process of image sampling and the limited sample size of UAV labeled images limit the accuracy of forest fire recognition based on UAV images. This paper proposes a FT-ResNet50 model based on transfer learning. The model migrates the ResNet network trained on an ImageNet dataset and its initialization parameters into the target dataset of forest fire identification based on UAV images. Combined with the characteristics of the target data set, Adam and Mish functions are used to fine tune the three convolution blocks of ResNet, and focal loss function and network structure parameters are added to optimize the ResNet network, to extract more effectively deep semantic information from fire images. The experimental results show that compared with baseline models, FT-ResNet50 achieved better accuracy in forest fire identification. The recognition accuracy of the FT-ResNet50 model was 79.48%; 3.87% higher than ResNet50 and 6.22% higher than VGG16.

Fire Detection Method in Smart City Environments Using a Deep-Learning-Based Approach

In the construction of new smart cities, traditional fire-detection systems can be replaced with vision-based systems to establish fire safety in society using emerging technologies, such as digital cameras, computer vision, artificial intelligence, and deep learning. In this study, we developed a fire detector that accurately detects even small sparks and sounds an alarm within 8 s of a fire outbreak. A novel convolutional neural network was developed to detect fire regions using an enhanced You Only Look Once (YOLO) v4network. Based on the improved YOLOv4 algorithm, we adapted the network to operate on the Banana Pi M3 board using only three layers. Initially, we examined the originalYOLOv4 approach to determine the accuracy of predictions of candidate fire regions. However, the anticipated results were not observed after several experiments involving this approach to detect fire accidents. We improved the traditional YOLOv4 network by increasing the size of the training dataset based on data augmentation techniques for the real-time monitoring of fire disasters. By modifying the network structure through automatic color augmentation, reducing parameters, etc., the proposed method successfully detected and notified the incidence of disastrous fires with a high speed and accuracy in different weather environments—sunny or cloudy, day or night. Experimental results revealed that the proposed method can be used successfully for the protection of smart cities and in monitoring fires in urban areas. Finally, we compared the performance of our method with that of recently reported fire-detection approaches employing widely used performance matrices to test the fire classification results achieved.

GIS-Based Forest Fire Risk Model: A Case Study in Laoshan National Forest Park, Nanjing

Fire risk prediction is significant for fire prevention and fire resource allocation. Fire risk maps are effective methods for quantifying regional fire risk. Laoshan National Forest Park has many precious natural resources and tourist attractions, but there is no fire risk assessment model. This paper aims to construct the forest fire risk map for Nanjing Laoshan National Forest Park. The forest fire risk model is constructed by factors (altitude, aspect, topographic wetness index, slope, distance to roads and populated areas, normalized difference vegetation index, and temperature) which have a great influence on the probability of inducing fire in Laoshan. Since the importance of factors in different study areas is inconsistent, it is necessary to calculate the significance of each factor of Laoshan. After the significance calculation is completed, the fire risk model of Laoshan can be obtained. Then, the fire risk map can be plotted based on the model. This fire risk map can clarify the fire risk level of each part of the study area, with 16.97% extremely low risk, 48.32% low risk, 17.35% moderate risk, 12.74% high risk and 4.62% extremely high risk, and it is compared with the data of MODIS fire anomaly point. The result shows that the accuracy of the risk map is 76.65%.

A Forest Fire Detection System Based on Ensemble Learning

Due to the various shapes, textures, and colors of fires, forest fire detection is a challenging task. The traditional image processing method relies heavily on manmade features, which is not universally applicable to all forest scenarios. In order to solve this problem, the deep learning technology is applied to learn and extract features of forest fires adaptively. However, the limited learning and perception ability of individual learners is not sufficient to make them perform well in complex tasks. Furthermore, learners tend to focus too much on local information, namely ground truth, but ignore global information, which may lead to false positives. In this paper, a novel ensemble learning method is proposed to detect forest fires in different scenarios. Firstly, two individual learners Yolov5 and EfficientDet are integrated to accomplish fire detection process. Secondly, another individual learner EfficientNet is responsible for learning global information to avoid false positives. Finally, detection results are made based on the decisions of three learners. Experiments on our dataset show that the proposed method improves detection performance by 2.5% to 10.9%, and decreases false positives by 51.3%, without any extra latency.

Integrating multiple factors to optimize watchtower deployment for wildfire detection

Traditional human-vision-based watchtower systems are being gradually replaced by the machine-vision-based watchtower system. The visual range of machine-vision-based watchtower is smaller than the range of traditional human-vision-based watchtower, which has led to a sharp increase in the number of towers that should be deployed. Consequently, the overlapping area between watchtowers is larger and overlaps are more frequent than in conventional watchtower networks. This poses an urgent challenge: identifying the optimal locations for deployment. If the number of required watchtowers must be increased to extend the detection coverage, overlaps among watchtowers are inevitable and result in viewshed redundancy. However, this redundancy of the viewshed resources of the watchtowers has not been utilized in the design of fire detection systems. Moreover, fire ignition factors, such as climatic factors, fuels, and human behaviour, cause the fire occurrence risk to differ among forest areas. Thus, the fire risk map of the area should also be considered in watchtower deployment. A fire risk model is used as the first step in producing the fire risk map, which is used to propose a new watchtower deployment model for optimizing the watchtower system by integrating viewshed analysis, location allocation, and multi-coverage of the high-fire-risk area while considering the budget constraints for providing optimal coverage. We use a real dataset of a forest park to evaluate the applicability of our approach. The proposed approach is evaluated against the FV-NB (Full coVerage with No Budget constraint) algorithm and the XV-B (maXimum possible coVerage with a Budget constraint) algorithm in terms of performance. The evaluation results demonstrate that our approach realizes higher coverage gain and excellent multiple-coverage of the fire risk area by integrating the viewshed and the fire risk level into location allocation while satisfying requirements on the overall coverage and budget. The proposed approach is more suitable in the environments with moderate watchtower density, in which overlapping areas are frequent. It offers as much as 8.9-17.3% improvement of multiple-coverage of the high-fire-risk area.

Accuracy Assessment of Digital Elevation Models (DEMs): A Critical Review of Practices of the Past Three Decades

An analysis of almost 200 references has been carried out in order to obtain knowledge about the DEM (Digital Elevation Model) accuracy assessment methods applied in the last three decades. With regard to grid DEMs, 14 aspects related to the accuracy assessment processes have been analysed (DEM data source, data model, reference source for the evaluation, extension of the evaluation, applied models, etc.). In the references analysed, except in rare cases where an accuracy assessment standard has been followed, accuracy criteria and methods are usually established according to the premises established by the authors. Visual analyses and 3D analyses are few in number. The great majority of cases assess accuracy by means of point-type control elements, with the use of linear and surface elements very rare. Most cases still consider the normal model for errors (discrepancies), but analysis based on the data itself is making headway. Sample size and clear criteria for segmentation are still open issues. Almost 21% of cases analyse the accuracy in some derived parameter(s) or output, but no standardization exists for this purpose. Thus, there has been an improvement in accuracy assessment methods, but there are still many aspects that require the attention of researchers and professional associations or standardization bodies such as a common vocabulary, standardized assessment methods, methods for meta-quality assessment, and indices with an applied quality perspective, among others.