| Title |
Deep learning-based cargo recognition and classification method for automated loading process in large-scale logistics |
| Authors |
김선목(Sun-mok Kim) ; 이상덕(Sang-Duck Lee) ; 최정아(Jung-A Choi) ; 이기백(Ki-beak Lee) |
| DOI |
https://doi.org/10.5370/KIEE.2024.73.1.192 |
| Keywords |
Camera Vision; Deep Learning; Logistics Automation; Object Detection; Real-Time |
| Abstract |
Automation of logistics center has a significant impact on industrial productivity. Currently, bottlenecks are frequently occurring due to human intervention in the loading and unloading processes of logistics. Thus, research for logistics automation has been continuously proposed mainly focused on cargo recognition and classification. However, there have been some limitations such as limited variety of compatible cargo dimensions and difficulty in continuous cargo recognition. To solve this problem, we propose a comprehensive cargo recognition and classification process to calculate the volume of cargo continuously fed onto a conveyor belt and classify the type of cargo. The volume calculation accuracy and classification accuracy of the test cargoes were approximately 93% and 98%, respectively. In addition, in continuous cargo loading tests considering the actual logistics center environment, volume calculation accuracy and classification accuracy were approximately 87% and 100%, respectively. |