Improved small target flame detection algorithm based on attention module

Wen Peng Wang; Yuan Bo Shi; Xiao Feng Wang

doi:10.1117/12.2689355

31 July 2023 Improved small target flame detection algorithm based on attention module

Wen Peng Wang, Yuan Bo Shi, Xiao Feng Wang

Proceedings Volume 12747, Third International Conference on Optics and Image Processing (ICOIP 2023); 127470U (2023) https://doi.org/10.1117/12.2689355
Event: Third International Conference on Optics and Image Processing (ICOIP 2023), 2023, Hangzhou, China

Abstract

In the field of flame detection, general target detection algorithms are troubled by low precision and thus difficult to detect small targets. In view of this situation and considering that flame detection requires detection algorithms to be real-time and accurate, a flame detection algorithm based on the attention module is proposed to improve Yolov5 networks. First, in the original network add the lightweight channel attention module ECA, and the network structure is adjusted to increase the feature extraction capability, so as to help the model locate flame features more accurately. Second, the Neck structure is modified by adding a small target detection layer, and in combination with other layers in the feature fusion section, a feature extraction and fusion module is added to improve the small target detection capability and feature fusion capability. Finally, the localization loss function is improved to EIoU Loss to accelerate the model convergence speed while improving the prediction frame regression precision. Experimental results show that the improved flame detection algorithm improves the average detection precision by 7.8% over the Yolov5s model, and the detection speed reaches 39 frame/s at the same time. It is not only real-time and accurate, but also effectively improves the detection capability on small targets.

Citation Download Citation

Wen Peng Wang, Yuan Bo Shi, and Xiao Feng Wang "Improved small target flame detection algorithm based on attention module", Proc. SPIE 12747, Third International Conference on Optics and Image Processing (ICOIP 2023), 127470U (31 July 2023); https://doi.org/10.1117/12.2689355

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
9 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Flame

Target detection

Detection and tracking algorithms

Flame detectors

Small targets

Feature extraction

Feature fusion

Keywords/Phrases

Search In:

Publication Years