An Early Flame Identification Method Based on Edge Gradient Feature

Ouyang, Jianquan; Bu, Leping; Zhikai, Yang; Wang, Teng

doi:10.1109/imcec.2018.8469419

Cited by 8 publications

(6 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Celik et al [14] used a single Gaussian model for the exclusion of non-flame regions, and T. Han [15] used a background subtraction method for the segmentation of flame regions. O. Jineng [16], on the other hand, introduced a method of edge gradient extraction to the flame edge color information for regional extrema to improve detection precision. Despite the rapid development of traditional computer vision feature extraction methods for flame features, they still can't detect in real-time.…”

Section: Related Workmentioning

confidence: 99%

“…The feature information is aggregated in both directions to produce a pair of direction-aware feature maps. This helps the network model to locate the object of the flame target more accurately and helps to occupy the channel information relationship efficiently, which is very effective in many studies [16,17]. The improved C3 module is named C3-ca and the network structure is shown in Fig.…”

Section: Coordinate Attention Blocksmentioning

confidence: 99%

See 1 more Smart Citation

YOLOv5-CSF: an improved deep convolutional neural network for flame detection

et al. 2023

View full text Add to dashboard Cite

The occurrence of fire threatens people's lives and property damage. An effective way to reduce flame damage is to detect flames in video images and take appropriate measures to contain fires at an early stage of occurrence. To solve the problem of low detection precision of current image-based flame detection methods, we propose a high precision flame detector based on yolov5 improvement. Coordinate Attention Blocks are introduced in the backbone network to obtain both channel relationships and long-range dependencies with precise positional information and increase the feature expression of the backbone network. Swin Transformer Blocks are introduced in the neck to expand the receptive field of the network and improve the flame feature extraction; the Adaptive Spatial Feature Fusion module is introduced in the head network to enhance the multi-scale feature fusion of flame features and reduce false alarms. Compared with the yolov5l, the precision of the model is improved by 4.1%. Compared with other existing detectors, it achieves the best average precision of 66.8%.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Coordinate Attention Blocksmentioning

confidence: 99%

YOLOv5-CSF: an improved deep convolutional neural network for flame detection

et al. 2023

View full text Add to dashboard Cite

show abstract

“…In recent years, the flame image detection technology extracts every frame captured by the monitor, and uses the image recognition detection technology for operation processing, which can accurately extract the suspected fire disaster area from the image, and then use the flame feature criteria to identify whether the fire occurs. Some scholars use RGB color model to extract flame [1][2][3][4][5] and establish HSI (hue, saturation, brightness) model to identify flame. The RGB color model is used to process the flame image, but it cannot convert the pixel value into chroma and brightness.…”

Section: Introductionmentioning

confidence: 99%

Research on flame image processing of climbazole production line based on RGB+HSI color model

Gao

et al. 2023

Third Optics Frontier Conference (OFS 2023)

View full text Add to dashboard Cite

The existing fire alarm system has strict distance and installation requirements between the fire point and the detector, and is easy to be interfered by environmental factors. It is not suitable for places with large space and many interference factors such as Climbazole production line. This paper proposed a flame image detection technology based on RGB+HSI color model and the detection system is designed and developed. The experimental results show that the flame image detection system based on RGB+HSI color model has the better recognition efficiency, which meets the real-time and accuracy requirements for early flame image detection in Climbazole production line.

show abstract

“…Qiu, T. et al [6] proposed an adaptive canny edge detection algorithm for fire image processing. Ji-neng, O. et al [7] proposed an early flame detection method based on edge gradient features. Khalil, A. et al [8] proposed a fire detection method based on multi-color space and background modeling.…”

Section: Introductionmentioning

confidence: 99%

FireViT: An Adaptive Lightweight Backbone Network for Fire Detection

Shen,

Sun,

et al. 2023

Forests

View full text Add to dashboard Cite

Fire incidents pose a significant threat to human life and property security. Accurate fire detection plays a crucial role in promptly responding to fire outbreaks and ensuring the smooth execution of subsequent firefighting efforts. Fixed-size convolutions struggle to capture the irregular variations in smoke and flames that occur during fire incidents. In this paper, we introduce FireViT, an adaptive lightweight backbone network that combines a convolutional neural network (CNN) and transformer for fire detection. The FireViT we propose is an improved backbone network based on MobileViT. We name the lightweight module that combines deformable convolution with a transformer as th DeformViT block and compare multiple builds of this module. We introduce deformable convolution in order to better adapt to the irregularly varying smoke and flame in fire scenarios. In addition, we introduce an improved adaptive GELU activation function, AdaptGELU, to further enhance the performance of the network model. FireViT is compared with mainstream lightweight backbone networks in fire detection experiments on our self-made labeled fire natural light dataset and fire infrared dataset, and the experimental results show the advantages of FireViT as a backbone network for fire detection. On the fire natural light dataset, FireViT outperforms the PP-LCNet lightweight network backbone for fire target detection, with a 1.85% increase in mean Average Precision (mAP) and a 0.9 M reduction in the number of parameters. Additionally, compared to the lightweight network backbone MobileViT-XS, which similarly combines a CNN and transformer, FireViT achieves a 1.2% higher mAP while reducing the Giga-Floating Point Operations (GFLOPs) by 1.3. FireViT additionally demonstrates strong detection performance on the fire infrared dataset.

show abstract

An Early Flame Identification Method Based on Edge Gradient Feature

Cited by 8 publications

References 7 publications

YOLOv5-CSF: an improved deep convolutional neural network for flame detection

YOLOv5-CSF: an improved deep convolutional neural network for flame detection

Research on flame image processing of climbazole production line based on RGB+HSI color model

FireViT: An Adaptive Lightweight Backbone Network for Fire Detection

Contact Info

Product

Resources

About