Automation of Sunspot Detection and Image Transmission by Machine Vision Processing at the Observatory of the Geophysical Institute of Peru, Province of Chupaca: Design and systemic modeling

This paper presents the design and implementation of an autonomous sunspot detection and transmission system developed at the observatory of the Geophysical Institute of Peru, in the province of Chupaca. Given the need for real-time solar monitoring systems at national observatories, the first stage involved developing an artificial vision model based on YOLOv7 together with the ODROID H4 Ultra processor. This computational model provides 70% efficiency compared to manual systems, calculated using performance metrics. In the second stage, the solar telescope was equipped with updated solar filters to enhance image recording. The overall operational performance of the system was largely positive, due to the accuracy of detection and ease of data transfer provided by the current processors. Validations between national or international observatories are recommended. Finally, the continuous improvement of the system developed for astronomical and research purposes is considered promising.