Digital And Statistical Analysis Of Hand-drawing Sunspot Images From The Purple Mountain Observatory

The Purple Mountain Observatory collected hand-painted sunspot information from five stations,which are the Beijing Observatory(referred to as BJP),Seshan Observatory(referred to as SSGCZ),the Purple Mountain Observatory(referred to as PMO),Qingdao Observatory(referred to as QDGX),and Yunnan Observatory(referred to as YNAO,partial year)have accumulated more than 35,000 hand-drawing sunspot images.Sunspot information was recorded on specific printing paper by the staff everyday,however,the sunspot data record was not digitized.Better to integrate and share the sunspot data between the observatories,hand-drawing sunspot images of five stations are digitized.The systematic difference analysis of the PMO data and from the perspective of visibility,PMO sunspot data observation quality was analyzed in the first time,which lay a solid foundation for the integration of the PMO data into the Solar Influences Data Analysics(SIDC)database.The article describes the digital process of the hand-paint sunspot images by the Purple Mountain Observatory,we compared relative sunspot number(PRSN)and grou p sunspot number(PGSN)of the PMO(data was missing in 1983 and 1984)with the corre sponding data(IRSN and IGSN)of SIDC from systematic difference and visibility.Mainly it was divided into the following three points:The Lab color space was used as the background extraction of the sunspot hand-drawing images.Firstly,the images were converted from RGB space to Lab color space,then a ?b color component were K-means cluster.The background area and the handwritten area were accurately divided into two categories,and then the nearest centroid classifier was used to extract the background of other hand-drawing pictures according to the clustering result.The extracted data was stored in a standardized format.The format of the digitized data was unified and the unified data was verified by the law of the sunspot number and the relationship between the sunspot number and the group sunspot number.In addition,it can also be verified according to the data range of each attribute,and the sun data map can be drawn by using the verified data,for example,the sunspot number change map,the sunspot butterfly chart,and the like.Unitary linear regression method was used to analyze relationship between the PMO data and SIDC data in stages.The relative sunspot number and group sunspot number of SIDC were used as dependent variables,and the corresponding the PMO data was used as an independent variable.The relative sunspot number and group sunspot number of PMO data and the SIDC data are analyzed systematic differences from two aspects.First the ratio changes of the PMO data and the SIDC data with time;second,the PMO data and the SIDC data change with time.The visibility of the PMO was divided into three grades:superior,medium and inferior,the effect of the PMO data was analyzed with changes of the visibility.Then R/S analysis was used to reveal the long-term persistence of sunspot activity time series from a statistical perspective,and the Lmob-Scargle periodogram method was used to calculate the solar activity cycle reflected by sunspot data in the observation time.