Stock taking activity is as a routine product inspection activity to check the inventory accuracy and help reduce the risks of stealing, damage, and obsolete inventories. This activity can be categorized as time consuming and expensive activity. In addition, this activity needs a lot of concentration and prone to human errors and mistakes. This study aims to replace human manual inspection in terms of object type and quantity with objects identification to reduce errors, time, and costs. Digital image processing in the form of Object Recognition is used in this study to determine the type of object and the number of objects. The results showed that the detection rate of a single product reached 90% which was influenced by the angle of an image and the detection rate of object quantity reaches 81% in average in real environment with a certain condition. It is expected that costs of inventory inspection and warehousing activities can be reduced, as well as the improvement in terms of efficiency and effectiveness can be achieved.
