Exploration on automatic identification algorithm of transmission line mountain fire based on image recognition technology

• Published in: International Journal of Thermofluids Vol. 20; p. 100494
• Main Authors: He, Wei, Chen, Wenjun, Wang, Yulin, Liu, Yu, Wang, Shenghong
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2023; Elsevier
• Subjects: Automatic recognition algorithm; Image feature extraction; Image recognition technology; Satellite meteorological data; Transmission line; Satellite meteorological data; Automatic recognition algorithm; Transmission line; Image feature extraction; Image recognition technology
• ISSN: 2666-2027; 2666-2027
• DOI: 10.1016/j.ijft.2023.100494

In recent years, many places have experienced frequent mountain fires, which have become one of the main disasters in the operation of power transmission lines. However, traditional manual inspection and video monitoring methods can only detect a small amount of mountain fires, and require a large amount of manpower and material resources. In this paper, image recognition technology was used to study the automatic identification algorithm of transmission line mountain fires, and image recognition technology was used to denoise the extracted images. After that, feature extraction was performed on the successfully denoised image, and the image was enhanced to improve the clarity of the image and prepare for improving the recognition accuracy of mountain fires. Through experiments, it can be found that using image recognition technology to identify mountain fires not only has high accuracy and recognition speed, but also has a lower error rate compared to using satellite meteorological data. The recognition accuracy of image recognition technology was above 95 %, while the recognition accuracy of using satellite meteorological data was below 92 %.