Establishing an Automatic System for Identifying the Stems of Salted Plums

Abstract. The main method for processing plums involves producing salted plums by salting fresh plums. After the stones are removed, the salted plums are processed into preserves. However, removing the stones of salted plums is a time-consuming task. The current method for removing stones entails manually loading the salted plums onto a pitting platform. Operating speeds are slow and must be improved. Pitting efficiency and yield rates are higher when the salted plums are positioned with their stems pointing down. Hence, the objectives of this study were to analyze the shapes of salted plums and identify the direction of stem points by using machine vision. When combined, these processes could identify stem positions automatically. In the future, the proposed automatic pitting system could be paired with automatic feeding devices to load salted plums with their stems pointing down on pitting platforms. This study divided the shapes of salted plums into four classes: tapered, blocked, rounded, and irregular. An analysis determined the ratios of each shape, with 63.4% being tapered plums, 17.1% blocked, 9.3% rounded, and 10.2% irregular. The tested results showed that the identification success rate for the tapered salted plums was the highest and reached 97%. The angles of the pointed tops on the tapered salted plums considerably facilitated identifying their shape. The overall success rate of the automatic shape identification was 88.5%. In the future, the system could be applied to continuous processors for ensuring consistently pitted salted plums, increasing the level of automation and operational effectiveness of processors used for pitting salted plums.