Video analytics for supply chain modelling – PostNL

 Overview

Company Description

One of PostNL’s most important service is to transport parcels ordered at web shops to consumers. To enable a sustainable growth PostNL is involved in a digital transformation of the supply chain, where capacity planning and steering is endorsed by both historic and realtime data sources. This transformation poses many challenges in the field of operational research and data science.

Project Description

At PostNL, we increasingly want to use video analytics to monitor and quantify process steps in our supply chain. For example, of every package a picture is taken and cameras will be installed at strategic points, e.g. during the loading and unloading process of parcels at the conveyer belts. We want to develop algorithms that are able to determine the size of a package, the moment of entry and exit in real time, and we want to link this data to data from other sensors, such as the blue-tooth beacons with which all our roll containers are equipped.

Goals of the Project
Explore methods to extract quantitative information from pictures and video data, identify the most promising methods and develop prototypes.

Deliverables

• Python code for video analytics
• Recommendations for further research based on given data and own explorations.

Essential Student Knowledge 

Python, image processing, deep learning, machine learning.

More information: escf@tue.nl