IMPLEMENTATION OF R-CNN IN OBJECT DIMENSION SURFACE AREA IDENTIFICATION AND ESTIMATION FOR STORAGE AREA EFFICIENCY IN OPEN WAREHOUSE

• Main Author: Ardiseno, Luqman
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/83638
• Institution: Institut Teknologi Bandung

The heavy equipment assembly industry faces challenges in object placement in open warehouses. This challenge becomes more difficult when objects are detected in various forms and the capacity of the open warehouse area is limited. The composition of storage in an open warehouse includes product components, raw materials, supporting components, work-in-progress, and waste. This research develops an open warehouse storage system using Region-based Convolutional Neural Network (R-CNN) algorithm. The system is modeled in an open warehouse prototype with the stages of identification and estimation of object dimensions. R-CNN identifies areas in the prototype open warehouse and estimates the dimensions of the detected objects. The obtained object dimensions are used to calculate the storage capacity of the area and provide recommendations for the layout of objects in the warehouse area. Evaluation on the open warehouse prototype shows that the system is able to identify and estimate object dimensions with good accuracy. This object identification can improve object tracking speed and make mobilization more effective through mapping, picking, moving, tracking, and object layout recommendations. The system has the potential to improve operator performance, facilitate industrial supervision, and increase storage area efficiency. Keywords: open warehouse, R-CNN, object identification, dimension estimation, laying recommendation, mobilization.