Deep Convolutional Generative Adversarial Networks for Flame Detection in Video

Aslan, Süleyman; Güdükbay, Uğur; Töreyi̇n, Behçet Uğur; Çetin, A. Enis

doi:10.1007/978-3-030-63007-2_63

Cited by 22 publications

(7 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the rapid development of computer vision, flame recognition research based on deep learning has received wide attention from scholars from all walks of life. Literature [12][13][14][15][16] researched convolutional neural networks for flame awareness detection with good recognition effect, but most of these methods use forest fires or artificial simulated flames as data set samples, while in the UAV aerial photography of grassland fires, there are differences between grassland fires and forest fires in terms of burning intensity, burning shape, etc. Therefore, this paper combines the characteristics of grassland fire spread and proposes a smoke recognition and dynamic grassland fire identification method.…”

Section: Introductionmentioning

confidence: 99%

Research on dynamic and static grassland fire identification based on UAV images

Zhang,

Liu,

Zhang

et al. 2023

Sixth International Conference on Computer Information Science and Application Technology (CISAT 2023)

View full text Add to dashboard Cite

China's grassland area is vast, once a grassland fire occurs, it will cause serious threat to people's life and property. At present, the domestic grassland fire monitoring means is mainly manual monitoring, this monitoring means of monitoring cost is large, low efficiency, real-time poor, and it is difficult to achieve complete detection, so the use of satellite, UAV images and other advanced means to identify grassland fire is the current research hot spot. In this paper, using UAV images, the detection is divided into two stages according to the burning process of grassland fires. In the early stage, the improved YOLOv5 network model is used to identify smoke due to the more obvious smoke characteristics; in the middle and late stage, the color characteristics of the fire area change obviously, and a flame detection method combining static and dynamic features is proposed. The C3 structure of the Neck part of the YOLOv5 model is improved, and the CBAM attention mechanism is introduced to make the model pay more attention to the target information; the loss function of YOLOv5 is improved, and DIoU is adopted as its loss function, and the experimental results show that the mAP of the improved YOLOv5 model is improved by 2.53% when trained on the self-built smoke dataset, and the recognition accuracy is effectively improved. When flame recognition is performed, the flame is affected by wind, vegetation coverage, humidity and terrain, etc., and spreads in all directions at different speeds, forming a dynamic outward-expanding circle with an increasing area and a distinct black color in the overfire area. Static color features are firstly used for pre-determination, and then the dynamic area and contour changes of the overfire are further discriminated. Experiments show that the algorithm reduces the false recognition rate of grassland fires and has good practicality.

show abstract

Section: Introductionmentioning

confidence: 99%

Research on dynamic and static grassland fire identification based on UAV images

Zhang,

Liu,

Zhang

et al. 2023

Sixth International Conference on Computer Information Science and Application Technology (CISAT 2023)

View full text Add to dashboard Cite

show abstract

“…In the past two decades, due to the rapid development of deep learning with good generalization and high degree of adaptation, and the performance improvement of graphics processing computing chips, various flame target detection models based on deep learning have emerged. Ko [2] et al proposed a flame detection model based on fuzzy finite automata, and Glan et al [3] proposed a deep convolutional generative adversarial neural network flame detection model based on vision. However, the parameters of these models are too large, resulting in insufficient computing power on edge platforms and limited deployment of pyrotechnic detection systems on them.…”

Section: Introductionmentioning

confidence: 99%

Research on lightweight detection model of spontaneous combustion of electric vehicles

Lu,

Yang,

Zhou

2023

Third International Conference on Advanced Algorithms and Signal Image Processing (AASIP 2023)

View full text Add to dashboard Cite

The article addresses the problems of large models, high cost, high false detection rate, low detection accuracy, and insufficient computing power of edge computing platforms that are common in current smoke and fire detection systems for parking lot fire monitoring. In this paper, we propose a lightweight self-ignition detection improvement model that is easy to use on monitoring devices and edge computing platforms. The model uses the YOLOv5 model as the base network, introduces an adaptive co-concern module, enhances the smoke and fire feature extraction capability of the backbone network, and uses sub-pixel sampling for semantic enhancement of the depth feature map. The size of the improved model is only 3.8MB, which can be more lightly deployed in edge surveillance environment for monitoring. The simulation comparison experiments before and after the model improvement are conducted in the self-combustion dataset under the self-built parking lot surveillance scenario with 3977 images in the training set and 896 images in the test set. The experimental results show that the error rate of the improved model for self-combustion smoke and fire detection is 8.1%, which is 2.7% lower than the error detection rate before the improvement; the detection accuracy reaches 82.0%, which is 5.2% higher than the accuracy before the improvement.

show abstract

“…[22] combined colour‐based rough extraction of smoke areas with CNN for flame detection; Aslan et al. [23] used deep convolution generative adversarial network (DCGAN) to realize flame detection based on supervised training. In order to improve the accuracy of detection, researchers put forward a new idea, which combined traditional image processing methods with depth neural networks, and achieved better results than a single method.…”

Section: Introductionmentioning

confidence: 99%

A hybrid method of detecting flame from video stream

Dou

Xie

et al. 2022

IET Image Processing

View full text Add to dashboard Cite

In this paper, a method of detecting flame from video stream is proposed exploiting the characteristics of the disordered movement, rapid deformation and intense colour of the flame. Firstly, the frame difference between video frame and background frame is calculated to obtain the main part of the moving object, and the difference between frames is calculated frame by frame in time series to obtain the deformation part of the moving object, and then the sum of cumulative difference between frames and the background difference between frames are added to generate a binary image containing the moving object and the deformed part. Secondly, the binary image is morphologically opened, and rectangular segmentation is carried out to obtain multiple suspicious flame regions. Finally, in the light of the intense colour of the flame, the corresponding area is extracted from the original picture by using the segmentation rectangle, and the colour statistics of the area are carried out to further judge whether there is a burning flame in the area. The experimental results show that the algorithm can accurately detect the burning area of flame in the real scene and eliminate the light interference and the movement interference.

show abstract

Deep Convolutional Generative Adversarial Networks for Flame Detection in Video

Cited by 22 publications

References 13 publications

Research on dynamic and static grassland fire identification based on UAV images

Research on dynamic and static grassland fire identification based on UAV images

Research on lightweight detection model of spontaneous combustion of electric vehicles

A hybrid method of detecting flame from video stream

Contact Info

Product

Resources

About