Customer

For more than 90 years, COWI has been among Denmark’s leading consulting companies specializing in engineering, the environment and social economy. They have offices all over the world, around 7,200 employees and always working on 12,000+ projects.

A special focus area is professional advice in connection with e.g. construction projects investigation of traffic conditions etc.; including securing the necessary analysis and documentation of existing conditions, development and results.

It was against this background and as a development of their service that COWI launched DDG in 2016, which today is available as an “on demand” service with HD panoramas of the Danish road network for professional use with image material collected by COWI’s cars.

COWI’s work with DDG today means that the whole of Denmark is photographed every two years. Every street, road, avenue and boulevard – all 108,000 km of the road network – is documented. In total, this adds up to almost 150 single or 30 million panoramic images in HD quality.

Challenge

GDPR and more

In relation to the GDPR, the specific task was that people as well as car number plates must be anonymised. You must not be able to identify who or which car has been captured by the photographer in a photo. Both faces and number plates must therefore be blurred.

It represented two sub-tasks: firstly, one must be able to determine whether it is a face or a number plate at all, and secondly, the blurring process must be able to be implemented efficiently, quickly and with an acceptable result.

Higher up – lower price

Basically, COWI had a desire to upgrade the existing DDG set-up and get a more flexible, scalable and cost-effective solution, where the workflow for processing the image material was at the same time automated to a much greater extent.

SOLUTION

KeyCore was chosen as a partner for the design of a cloud infrastructure in which the processing of image material could take place more cost-effectively and scalable compared to the existing on-premise solution.

To handle this, processing of the imagery for DDG was migrated to an AWS cloud infrastructure.

The setup of COWI’s cloud infrastructure was orchestrated taking into account the strict guidelines of the GDPR, so that storage falls into two stages. According to the GDPR, it is not allowed to store non-anonymized image material of this nature for more than six months. It was therefore of vital importance that the material could automatically go through the anonymisation process immediately after uploading.

With AWS S3 Standard, COWI gained access to be able to turn up to maximum capacity during the short period that image material was being processed. At the same time, S3 has automated backup, which provides a guarantee for 99.999999999% of the material on an annual basis, or a maximum loss corresponding to one object in 10,000,000 in 10,000 years.

The long-term storage is handled in an AWS S3 Glacier – which costs only $0.004 per gigabytes per month – and AWS Batch, which both ensures scalability and access to resources and speed, as well as that more resources are never seized than necessary and that they are used when the price is lowest.

Security

By migrating to the AWS cloud, COWI also gained access to the highest imaginable level of compliance for everything below the hypervisor line. It includes infrastructure, hardware, the physical facilities and networks. It is checked and documented several times a year by a third party so that users of AWS can use it as part of their documentation showing that they are compliant and so that they can obtain the necessary certification that may apply to them. It is compliant with, among other things, ISO 27001, GDPR, EU-US Privacy Shield, HIPAA and HMDA.

Post processing

Part of the post-processing of the recordings had previously been done manually, which not only entailed a large consumption of resources, but also increased the risk of errors.

With KeyCore’s infrastructure in the cloud, COWI was able to ensure that the entire process was automated and streamlined using ML.

Specifically, KeyCore’s architecture gave COWI the opportunity to use the market’s absolutely leading algorithms in an extremely modular container-based setup, where COWI can now very simply replace modules in step with the development of the models. Among other things, several different Deep Learning Convolution Neural Networks (CNN), where you first “train” neural networks on already classified images and can subsequently classify and identify new material.

On the cloud infrastructure that KeyCore has developed for COWI, the post-processing has become far more automated, which ensures fast and efficient post-processing from when the image material is uploaded until it is ready for publication in DDG. In addition to ensuring the GDPR-required blurring of faces and number plates, the cloud processing also includes stitching of the panoramic images, correction of the image quality and a dashboard for monitoring the status of the processing of uploaded material that gives COWI’s customers a valuable tool in their work every day.

Scroll to Top