Checklist: Image Blurring for Mobile Mapping

A list of things you need to consider when starting with the blurring of panorama and planar images to comply with GDPR and other data privacy laws.

08 June 2020, by Mario Sabatino Riontino

Figure 1: Blurred faces. Eiffel Tower, Paris (Bing Maps Streetside)
Figure 1: Blurred faces. Eiffel Tower, Paris (Bing Maps Streetside)

Why image blurring is important

As we discussed already in a previous article, face and body are the most fundamental and highly visible elements of our identity. Hence, they fall under the definition of personally identifiable information (PII) of the General Data Protection Regulation (GDPR).

That said, when using or publishing an image or video for commercial purposes, the regulation requires the data processor (e.g. companies) to request consent from the data subject, i.e. the person whom the PII belongs to.

Consent expresses the voluntary authorization (in written form) to process the requested data.

Unfortunately, when you consider large image datasets, getting consent will be time-consuming and costly due to the hundreds of thousands, or even millions of individuals within the dataset.

That’s when anonymization comes into play, making the data subject no longer identifiable. In the specific case of image processing, blurring emerged as the de-facto standard anonymization method used by companies like Google, Microsoft and Apple.

The checklist

✓ What needs to be anonymized?

Typically, faces and license plates are the most requested object types when it comes to blurring. But your customer might also demand to blur whole bodies and vehicles. When it comes to indoor mapping, blurring computer screens, whiteboards or door signs could also be a requirement.

✓ How many images need to be anonymized?

The amount of images impacts costs and processing time. To make the data transfer less error-prone, larger amounts of data might need to be chunked up into smaller packages.

✓ Is anonymization time-critical?

Project deadlines sometimes necessitate the image blurring to be as fast as possible. Cloud-based anonymization solutions are highly scalable, i.e. multiple computing nodes can work on the anonymization in parallel, and thus the preferred method. On-premise software cannot be scaled without purchasing and maintaining additional hardware.

Total time = (amount of images * processing time per image) / computing nodes

✓ Use a Cloud or On-premise Anonymization Solution?

Here's a comparison of the advantages and disadvantages:

Cloud On-premise
Costs ❌ Slightly higher price per image
✅ No costs of ownership
✅ Slightly lower price per image
❌ Higher total costs of ownership (hardware, software, trained personnel)
Scalability (Speed) ✅ Highly scalable ❌ Hard to scale
Maintenance ✅ Maintenance-free ❌ Needs to be maintained
Idle time ✅ No idle time ❌ Idle time
Accessibility ✅ Accessible from nearly everywhere ❌ Accessibility might be limited
✅ Transfer of large amounts of data might be easier
Data security & privacy ✅ GDPR-compliant data centers ✅ Data remains within your own data center
❌ You need to have data protection measures in place

✓ What type of images should be anonymized?

Image blurring can be applied to panorama as well as planar images. Depending on your use case, only one or both types need to be anonymized. For a viewer software similar to Google Street View, anonymizing panoramas alone will likely suffice. More advanced surveying software often allows the user to display planar in addition to panorama images as well, thus blurring for both image types might be necessary.

✓ Data privacy and security due diligence

Make sure the anonymization provider of your choice meets the following requirements:

  • Technical and organizational measures (TOM) to ensure data protection
  • Records of Processing Activities
  • Dedicated Data Protection Officer (DPO) at service
  • Data Protection Audits (recurring)
  • Processing takes place in GDPR-certified data centers
  • All data and storage devices are encrypted

✓ Use an automated blurring solution, instead of manual labor

Relying on companies which provide manual image blurring comes with certain problems and risks. Manual labour does not scale, thus only a limited amount of images can be blurred in a given time. Companies located outside of your home country may have different, weaker data protection laws. Each person that comes in contact with your imagery increases the risk of a security breach.

Automated solutions reduce those risks and problems. Cloud-based solutions are scalable and based in e.g. GDPR-compliant data centers in the EU, also the number of persons who interact with your data is drastically reduced.

Celantur’s solution

Celantur offers a solution to automatically anonymize large amounts of images and videos. We enable our customers to blur personal data like faces, bodies, license plates and vehicles.

Our technology uses convolutional neural networks (CNN) for object detection in images and videos, which in recent years has rapidly evolved from an academic playground to production-ready tools for solving industrial use-cases fast and effectively. Our solution is available on our cloud platform, as well as an on-premise software.

Since Data Protection is at the core of our business, we have strong measures in place to comply with the GDPR and other data protection laws. Up-to-date documentation is available to our customers for legal compliance.


Getting started with image blurring for mobile mapping isn’t rocket science. By following our checklist, you’ll cover the most relevant steps before choosing a solution. If you have any further questions or want to get a free demo of our state-of-the-art image blurring solution, feel free to contact us any time.

This is the second post in our series about mobile mapping. Feel free to check out the first one about Data Protection for Mobile Mapping.

mobile mappingdata protectiongdprenglish
Contact Us

Latest Blog Posts

Checklist: Anonymization for ADAS and Autonomous Vehicles

All you need to know to anonymize ADAS datasets and comply with data protection laws

Drones, UAV and Data Protection in the EU

Learn how to deal with the sensitive issue of data privacy in the European Union when planning a drone project.

Understanding Specificity and Sensitivity

How these two kind of wrong can impact your computer vision project for object detection.