Multi-parameter fire detection method based on feature depth extraction and stacking ensemble learning model

Qu, Na; Li, Zhongzhi; Li, Xiaoxue; Zhang, Shuai; Zheng, Tianfang

doi:10.1016/j.firesaf.2022.103541

Cited by 23 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The smoke from the real fires used in this study complies with European standard fires, such as beech wood smoldering fire (TF2), cotton smoldering fire (TF3), polyurethane open flame (TF4), and n-heptane open flame (TF5) [ 19 , 24 , 25 , 26 , 27 , 28 ]. Moreover, a survey by the National Fire Protection Association (NFPA) reported that the most likely cause of false alarms by detectors is oil fumes because their particle sizes as well as their refractive indices being very close to that of real fire smoke [ 29 ].…”

Section: Methodsmentioning

confidence: 99%

“…This method is impractical for commercial photoelectric detectors due to the large size and complexity of the light intensity measurement device and the requirement of scattering information in the deep UV for classification. Qu et al [ 19 ] classified four classes of European standard fires and typical interfering aerosols using a combination of multiple parameters, such as temperature, smoke, and CO concentration. Similarly, Yu et al [ 20 ] proposed multi-detector, real-time fire alarm technology to classify oil fumes and multiple types of real fire smoke.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Embedded Spatial–Temporal Convolutional Neural Network Based on Scattered Light Signals for Fire and Interferential Aerosol Classification

Xu,

Zhu,

Lin

et al. 2024

Sensors

View full text Add to dashboard Cite

Photoelectric smoke detectors are the most cost-effective devices for very early warning fire alarms. However, due to the different light intensity response values of different kinds of fire smoke and interference from interferential aerosols, they have a high false-alarm rate, which limits their popularity in Chinese homes. To address these issues, an embedded spatial–temporal convolutional neural network (EST-CNN) model is proposed for real fire smoke identification and aerosol (fire smoke and interferential aerosols) classification. The EST-CNN consists of three modules, including information fusion, scattering feature extraction, and aerosol classification. Moreover, a two-dimensional spatial–temporal scattering (2D-TS) matrix is designed to fuse the scattered light intensities in different channels and adjacent time slices, which is the output of the information fusion module and the input for the scattering feature extraction module. The EST-CNN is trained and tested with experimental data measured on an established fire test platform using the developed dual-wavelength dual-angle photoelectric smoke detector. The optimal network parameters were selected through extensive experiments, resulting in an average classification accuracy of 98.96% for different aerosols, with only 67 kB network parameters. The experimental results demonstrate the feasibility of installing the designed EST-CNN model directly in existing commercial photoelectric smoke detectors to realize aerosol classification.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Embedded Spatial–Temporal Convolutional Neural Network Based on Scattered Light Signals for Fire and Interferential Aerosol Classification

Xu,

Zhu,

Lin

et al. 2024

Sensors

View full text Add to dashboard Cite

show abstract

“…Damage caused by fire is becoming more serious increasingly with the continuous development of social economy. If the fire is detected as soon as possible, the loss can be reduced by more than 80% (Qu et al 2022). Early detection of fire is very important to reduce the loss due to fire.…”

Section: Theoretical Backgroundmentioning

confidence: 99%

Comparing the performance of Genetic Algorithm and Particle Swarm Optimization Algorithm in allocating and scheduling fire stations for dispatching forces to a fire/accident (A Case study: the Region 19, Tehran, Iran)

Kheirdast,

Jozi,

Rezaian

et al. 2023

Preprint

View full text Add to dashboard Cite

Considering the importance of "time" in the process of dispatching forces to reach the fire or accident site, GA or PSO models can be used as artificial intelligence alternatives. Genetic Algorithm (GA) and Particle Swarm Optimization Algorithm (PSOA) models can be used. This research shows which of these two models is more appropriate in this case study. With the hypothesis that GA and PSOA have positive effects on the allocation and scheduling of the stations, this research seeks to compare them in order to find which one of these two methods is more appropriate to shorten the time to reach fire/incident site in the Region 19 of Tehran. This is an applied type of research. Data analysis was carried out using NFPA standards and MATLAB software. The statistical population includes 8 fire stations and 250 personnel of the stations selected in a simple way, and the sampling volume was obtained using Morgan's table (n = 148). At first, the algorithm of dispatching forces to reach the site of fire/incident was designed and implemented based on PSOA, GA and the time to response the incident according to NFPA1720 standards. After writing the assumptions of the problem and running the mathematical model from nonlinear to linear, the data was entered into the MATLAB software, and finally by comparing the performance improvement of PSOA and GA, appropriate results were obtained. In order to efficiently assign and schedule fire stations to arrive at the site, a linear numerical programming model was presented with the aim of minimizing the arrival time and taking into account the effect of firemen's fatigue (α = 0.1). The findings of the research showed that the operation processing time (of fire extinguishing) had a normal distribution with a mean of 40 minutes and a variance of 10 minutes, independent of the severity of the incident. Also, fatigue coefficient was calculated 0.1 by analyzing the sensitivity of the solution time of the algorithm with changes [0–1]. The initial standard travel time, with an average speed of 47 km/h and a density factor of 1.24, was 5min:20s. Solving the problem in large and small dimensions showed that the initial power effect of each fire station is 0.36 according to the fatigue level of the forces. Based on the obtained results, GA performs better in terms of problem solution time, and the improved PSOA also has higher quality answers.

show abstract

“…It has achieved groundbreaking success in a wide range of real-world applications [61]. Within the domain of deep learning, the 1D Convolutional Neural Network (1D-CNN) has garnered attention for its ability to capture meaningful features, making it suitable for both classification and regression tasks in environmental modeling and predictions, i.e., fires [62], landslides [63], and floods [24].…”

Section: D-convolution Neural Networkmentioning

confidence: 99%

Spatial Prediction of Fluvial Flood in High-Frequency Tropical Cyclone Area Using TensorFlow 1D-Convolution Neural Networks and Geospatial Data

Trong,

Quang,

Cuong

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

Fluvial floods endure as one of the most catastrophic weather-induced disasters worldwide, leading to numerous fatalities each year and significantly impacting socio-economic development and the environment. Hence, the research and development of new methods and algorithms focused on improving fluvial flood prediction and devising robust flood management strategies are essential. This study explores and assesses the potential application of 1D-Convolution Neural Networks (1D-CNN) for spatial prediction of fluvial flood in the Quang Nam province, a high-frequency tropical cyclone area in central Vietnam. To this end, a geospatial database with 4156 fluvial flood locations and 12 flood indicators was considered. The ADAM algorithm and the MSE loss function were used to train the 1D-CNN model, whereas popular performance metrics, such as Accuracy (Acc), Kappa, and AUC, were used to measure the performance. The results indicated remarkable performance by the 1D-CNN model, achieving high prediction accuracy with metrics such as Acc = 90.7%, Kappa = 0.814, and AUC = 0.963. Notably, the proposed 1D-CNN model outperformed benchmark models, including DeepNN, SVM, and LR. This achievement underscores the promise and innovation brought by 1D-CNN in the realm of susceptibility mapping for fluvial floods.

show abstract

Multi-parameter fire detection method based on feature depth extraction and stacking ensemble learning model

Cited by 23 publications

References 16 publications

Embedded Spatial–Temporal Convolutional Neural Network Based on Scattered Light Signals for Fire and Interferential Aerosol Classification

Embedded Spatial–Temporal Convolutional Neural Network Based on Scattered Light Signals for Fire and Interferential Aerosol Classification

Comparing the performance of Genetic Algorithm and Particle Swarm Optimization Algorithm in allocating and scheduling fire stations for dispatching forces to a fire/accident (A Case study: the Region 19, Tehran, Iran)

Spatial Prediction of Fluvial Flood in High-Frequency Tropical Cyclone Area Using TensorFlow 1D-Convolution Neural Networks and Geospatial Data

Contact Info

Product

Resources

About