Improving a Night-Time Weather Estimation Method with Low-Cost Devices using a Support Vector Machine
Abstract
Because of the increasing density of meteorological observation networks and technological advances in data processing, weather forecast accuracy has improved in the past decades and is projected to increase further, requiring measurements at the local scale. However, their biggest problem is the equipment cost; the ultrasensitive atmospheric light cameras cost approximately ten million yen. This paper introduces a weather observation system designed exclusively with low-cost commercial products. The combined cost of the fabricated device was approximately 100,000 yen. An algorithm was also established to estimate night-time weather from the number of visible stars in all-sky images by assuming that more visible stars imply fewer clouds and more precise weather (the star count method). The resulting weather estimation accuracy was about 80%. In addition, to improve the accuracy, this paper adopts a supervised machine learning method, Support Vector Machine (SVM), and compares the classification accuracy with the star count method.
References
Shin Suzuki. Development of a multi-station airglow imaging system using commercial cameras (in Japanese). In MGI34-P08, Japan Geoscience Union Meeting, 2021.
Shiokawa, Kazuo et al. Ground-based instruments of the PWING project to investigate dynamics of the inner magnetosphere at subauroral latitudes as a part of the ERG-ground coordinated observation network, url = https://doi.org/10.1186/s40623-017-0745-9. Earth, Planets and Space, 69(1):160, 2017.
Kazuo Shiokawa and Shin Suzuki. Wonder of the Airglow (in Japanese). Gentosha, 2018.
Shin Suzuki and Keiichiro Fukazawa. Development of Airglow Camera System for Multi-Point Airglow Imaging Measurements: Detection of Wave Structures (in Japanese). Aichi University Journal of Regional Policy Studies, 8(1·2):3–8, 2019.
Yasunobu Ogawa et al. Development of low-cost multi-wavelength imager system for studies of aurora and airglow. Polar Science, 23:100501, 12 2019.
Shin Suzuki and Kazuo Shiokawa and Alan Z. Liu and Yuichi Otsuka and Tadahiko Ogawa and Takuji Nakamura. Characteristics of equatorial gravity waves derived from mesospheric airglow imaging observations. Annales Geophysicae, 27(4):1625–1629, 2009.
Shota Kusabiraki and Shin Suzuki and Kazunori Iwata. Development of a Night-time Weather Estimation Algorithm Using Image Processing with Low-Cost Devices. In 2023 14th IIAI International Congress on Advanced Applied Informatics (IIAI-AAI), pages 271–276, Los Alamitos, CA, USA, Jul. 2023. IEEE Computer Society.
Japan Meteorological Agency. Weather overview terminology. https://www.data.jma.go.jp/obd/stats/data/mdrr/man/gaikyo.html (Accessed: 2023/12/4).
Ioannis Tsochantaridis, Thorsten Joachims, Thomas Hofmann, and Yasemin Altun. Large margin methods for structured and interdependent output variables. Journal of Machine Learning Research, 6(50):1453–1484, 2005.
Evidently AI. Accuracy, precision, and recall in multi-class classification. https://www.evidentlyai.com/classification-metrics/multi-class-metrics (Accessed: 2024/1/28).
