SunCounter: Software for automated reading of sunshine duration recorded by heliographs

Abstract

The aim of this study was to develop software for reading and managing sunshine duration recorded by Campbell–Stokes heliographs. The software employs digital image processing techniques, such as mean filtering, thresholding, and opening, to interpret digitized images of record cards containing sunshine duration measurements. The software's validation was conducted by comparing global solar irradiation measurements with estimates of global solar irradiation generated from sunshine duration obtained automatically by the software and manually by a meteorological observer. Estimates generated by the automated method showed better performance (mean bias error: 0.084, relative mean bias error: 0.500, root mean square error: 2.045, relative root mean square error: 12.109, and correlation coefficient: 0.954). Additionally, the automated method was significantly faster, taking an average of 41.825 s less than the manual method to perform the readings. This suggests that the software can be used to automate, standardize, and speed up reading sunshine duration data. Furthermore, the use of the software enabled the creation of a daily sunshine duration database, common in meteorological stations, and an hourly database, which is a novelty due to the complexity of the manual reading method.