정지궤도 환경위성과 인공지능을 이용한 산불 연기 탐지
Detection of Wildfire Smoke using GEMS Imagery and AI
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Wildfires have been occurring for a long time due to natural and anthropogenic causes, but recently the frequency and intensity of wildfires have increased significantly. Wildfires are difficult to identify after they occur, and once they occur, they cause serious damage to the environment and society. Accordingly, various studies are being conducted with the goal of rapid and effective wildfire detection. However, research on large-scale pollution estimation using satellite remote sensing is still insufficient. This study aims to generate wildfire smoke detection products using GEMS(Geostationary Environment Monitoring Spectrometer) sensors. First, candidate pixels for wildfires were determined through the Swin transformer model, which has recently attracted attention in the image recognition field. In the next step, over-detection areas were suppressed through a denoising model created with a random forest, and a yellow dust classification model was built to exclude the possibility of false detection due to yellow dust among the finally detected pixels to generate wildfire smoke detection products. The evaluation accuracy of each built model was mIoU=0.858 and F1 score=0.920 for the Swin transformer, mIoU=0.908 and F1 score=0.926 for the denoising model, and mIoU=0.993 for the wildfire smoke and yellow dust classification model.
Abstract
Wildfires have been occurring for a long time due to natural and anthropogenic causes, but recently the frequency and intensity of wildfires have increased significantly. Wildfires are difficult to identify after they occur, and once they occur, they cause serious damage to the environment and society. Accordingly, various studies are being conducted with the goal of rapid and effective wildfire detection. However, research on large-scale pollution estimation using satellite remote sensing is still insufficient. This study aims to generate wildfire smoke detection products using GEMS(Geostationary Environment Monitoring Spectrometer) sensors. First, candidate pixels for wildfires were determined through the Swin transformer model, which has recently attracted attention in the image recognition field. In the next step, over-detection areas were suppressed through a denoising model created with a random forest, and a yellow dust classification model was built to exclude the possibility of false detection due to yellow dust among the finally detected pixels to generate wildfire smoke detection products. The evaluation accuracy of each built model was mIoU=0.858 and F1 score=0.920 for the Swin transformer, mIoU=0.908 and F1 score=0.926 for the denoising model, and mIoU=0.993 for the wildfire smoke and yellow dust classification model.
참고문헌 (22)
- Alvarado, L., "Unexpected long-range transport of glyoxal and formaldehyde observed from the Copernicus Sentinel-5 Precursor satellite during the 2018 Canadian wildfires", Atmospheric Chemistry and Physics, 20(4), pp.2057~2072 (2020)
- Arrue, B. C., "An intelligent system for false alarm reduction in infrared forest-fire detection", IEEE Intelligent Systems and their Applications, 15(3), pp.64~73 (2000)
- Cai, W., "A real-time smoke detection model based on YOLO-SMOKE algorithm", 2020 Cross Strait Radio Science and Wireless Technology Conference (CSRSWTC), pp.1~3 (2020)
- Crutzen, P. J., "Biomass burning in the tropics: impact on atmospheric chemistry and biogeochemical cycles", Science, 250(4988), pp.1669~1678 (1990)
- Evans, L. F., "Photochemical ozone in smoke from prescribed burning of forests", Environmental Science and Technology, 11(9), pp.896~900 (1977)
- Grammalidis, N., "A multi-sensor network for the protection of cultural heritage", 2011 19th European Signal Processing Conference, pp.889~893 (2011)
- Günay, O., "Fire detection in video using LMS based active learning", Fire Technology, 46, pp.551~577 (2010)
- Hally, B., "Advances in active fire detection using a multi-temporal method for next-generation geostationary satellite data", International Journal of Digital Earth, 12(9), pp.1030~1045 (2018)
- He, X., "AutoML: a survey of the state-of-the-art", Knowledge-Based Systems, 212, pp.106622~ (2021)
- Hsu, N. C., "Detection of biomass burning smoke from TOMS measurements", Geophysical Research Letters, 23(7), pp.745~748 (1996)
- Jethva, H., "Satellite-based evidence of wavelength-dependent aerosol absorption in biomass burning smoke inferred from Ozone Monitoring Instrument", Atmospheric Chemistry and Physics, 11(20), pp.10541~10551 (2011)
- 김형우, "HRNet-OCR과 Swin-L 모델을 이용한 조식동물 서식지 수중영상의 의미론적 분할", 대한원격탐사학회지, 38(5), pp.913~924 (2022), 대한원격탐사학회[DOI]
- Krüger, F., "Activity, Context, and Plan Recognition with Computational Causal Behaviour Models", Doctor, pp.2016
- Larsen, A., "A deep learning approach to identify smoke plumes in satellite imagery in near-real time for health risk communication", Journal of Exposure Science and Environmental Epidemiology, 31(1), pp.170~176 (2021)
- Li, X., "Forest fire smoke detection using back-propagation neural network based on MODIS data", Remote Sensing, 7(4), pp.4473~4498 (2015)
- Liu, Z., "Swin transformer:hierarchical vision transformer using shifted windows", Proceedings of the IEEE, 2021, 10012~10022
- Moon, Y. H., "Recent research and development trends in automated machine learning", Electronics and Telecommunications Trends, 34(4), pp.32~42 (2019)
- Ogunjobi, K. O., "Aerosol optical depth during episodes of Asian dust storms and biomass burning at Kwangju, South Korea", Atmospheric Environment, 38(9), pp.1313~1323 (2004)
- 신성균, "Trend Analysis of Wildland Fires and Their Impacts on Atmospheric Environment over East Asia", Asian Journal of Atmospheric Environment, 10(1), pp.22~31 (2016), 한국대기환경학회[DOI]
- Torres, O., "Derivation of aerosol properties from satellite measurements of backscattered ultraviolet radiation: theoretical basis", Journal of Geophysical Research: Atmospheres, 103(D14), pp.17099~17110 (1998)
- Urbanski, S. P., "Chemical composition of wildland fire emissions", Developments in Environmental Science, 8, pp.79~107 (2008)
- Wang, Z., "Semantic segmentation and analysis on sensitive parameters of forest fire smoke using smoke-unet and landsat-8 imagery", Remote Sensing, 14(1), pp.45~ (2021)