Sustainable PRS: A hybrid DL-based parcel recognition approach

Parcel packaging and delivery has been an integral part of the logistic and e-commerce domain. With the high demand and supply of these parcel packages, there is a huge amount of mismanagement in the system which led to the damaging of parcel packages. Therefore the proposed work focus on developing a sustainable and efficient parcel recognition system (PRS) using the DL technique with the help of hybridization of CNN and LSTM. The complete experiment has been divided into two steps, first is to recognition of fine and faulty parcel packages resulting in an accuracy of 93.83%, and in the second step multi-classification of faulty parcel packages has been carried out utilising four separate tier levels of parcel faultiness which resulted into the best classification accuracy of 98.27% for level 2 of faulty parcel package. The complete study and proposed DL-based hybrid approach have proved to be an effective, efficient and sustainable technique for the recognition and classification of parcel packages to establish a sustainable PRS. In addition to that, the proposed work will contribute to a better quality of life, technology transfer, and community enhancement for both readers and researchers in the suggested domain.