AWS and Deloitte Help Port of Vancouver Improve Supply Chain Visibility with Computer Vision-based Container Tracking

In 2017, the Vancouver Fraser Port Authority (Port of Vancouver) identified a need to enhance coordination, transparency, and supply chain efficiency of the container examination process for the Western Canadian Gateway. It was a major source of complaints and challenges among the entire port community. Reliable information on actual operations – which are distributed across multiple organizations – was limited. 

The Port and Deloitte chose computer vision to help optimize ground operations and expedite container inspection by more effectively tracking movement of shipping containers subject to examination. Working with AWS, Deloitte aims to use AWS Panorama’s edge computer vision inference capability and other advanced analytics technologies to identify and track containers, feed real-time data into the blockchain system, enable proper assignment of cost, and identify efficiency improvement opportunities in the existing process.  

In 2017, the container examination supply chain for the Western Canada Gateway, a trade and logistics corridor, needed an overhaul. “It was a major source of complaints and challenges for the entire port community. Reliable information on actual container examination operations was limited, many processes were manual, and information was distributed across multiple organizations,” said Greg Rogge, Director of Land Operations and Security at the Port of Vancouver.

The variability and inefficiency of the process also affected costs and timelines. So, in 2018, in partnership with Deloitte, Port of Vancouver introduced a blockchain application to add visibility and transparency into the container examination process. After deploying the blockchain application, Port of Vancouver and Deloitte saw an opportunity to apply other advanced technologies to optimize container tracking and movement.

A field of AI, computer vision enables organizations to create new insights from their visual data. Using deep learning techniques, machines can accurately identify and classify objects in images and video – and then react to what they “see.” The Port of Vancouver and Deloitte decided to use AWS Panorama for its container tracking solution because it processes data locally, at the edge, rather than the cloud. Due to the distributed nature of the Port’s operations, bandwidth constraints, and requirements for data residency, AWS Panorama and its ability to run computer vision models locally stood out as a preferred solution. 

With AWS Panorama, Deloitte is helping the Port develop a new class of insights on its physical operations that will improve collaboration with supply chain partners. “We identified a novel approach – integrating AWS Panorama with the existing blockchain application,” said Don MacPherson, Partner, Deloitte. Implementing this approach would reduce manual data entry even more, improve data reconciliation, and provide more accurate, real-time, and rich information into the application for key stakeholders. An added benefit was the fact that Port of Vancouver could use its existing IP camera infrastructure as part of the solution. 

Home to 27 terminals, the Port has limited access to additional industrial land for expansion. With space for inspection at a premium, efficient ground operations are key. ​With Deloitte’s computer vision solution on AWS Panorama, improved container detection, tracking, and localization throughout the drayage and exam process has accelerated container inspection and streamlined traffic. A bot predicts and sends alerts about berth allocation, container sequencing, and containers ready to return, which will decrease wait time and use port space more efficiently.  

Edge computing limits the transmission of video data about containers, its contents, and conditions, much of which is private, over the internet. An on-premises edge server can process the data securely and return critical information needed for near real-time applications. Because the container examination process depends heavily on handheld devices, cameras, and IoT, AWS Panorama was a good fit for VFPA. “By adopting computer vision at the edge, Port of Vancouver is increasing its competiveness while preserving privacy,” said Rogge.

Inefficient and manual container examination processes are costly. Anything from missing paperwork and intensive exams can delay shipments from a few days to a month. Some require additional labor for unloading and reloading cargo. They interrupt traffic flow and ground operations and increase storage and labor costs. With the automation provided by the Deloitte and AWS solution, Port of Vancouver is reducing delays and costs. It is also improving the experience of container inspectors, truck drivers, terminal employees, and other port clients because the automation results in more productivity and fewer bottlenecks and delays at the port. “With AWS Panorama feeding real-time data into our existing blockchain system, computer vision can enable the proper assignment of cost and continue to identify efficiency improvement opportunities in the existing process,” said MacPherson.​