AeroScan Case Study
Aeroscan digitalizes building inspection and maintenance using photos and 3D models of the constructions. We use drones equipped with cameras to take photos of the buildings from various angles and distances. These photos are then used to create a digital 3D model of the construction. These models together with the photo sets are delivered to our clients on our online platform called Aespect.
What was the main challenge?
Besides the main object of interest, drone cameras also capture the whole surroundings, including bystanders and pedestrians. We are required to anonymize these images to be GDPR compliant. Pedestrians can be visible in photos at any distance and angle, which makes the anonymization problem more difficult for AI algorithms.
At the end of each of their working day, drone operators upload collected photo sets to our data processing servers. In the past, our employees would need to inspect the photos personally and blur out pedestrians themselves by hand. This was a tedious and time-consuming process.
Currently we run Gallio PRO on all photo sets after they are uploaded. This way, we anonymize the photos as early as possible.
What are the benefits of using software?
Gallio PRO provides us with two large advantages:
- the time period when Aeroscan has possession of unanonimyzed photos, has become a lot shorter.
- we save a lot of time an effort that otherwise would be spent on manual anonymization (and time is money, they say).
The team supported us with custom development services, to allow some extent of automation and better integration with our existing IT infrastructure.
The detection blurring quality of Gallio PRO is generally very good. However, we have noticed lower accuracy in extreme cases, for example when cameras take photos downwards at an angle of -70 degrees, which is common practice with drone photography. Fortunately, in such photos pedestrians are usually very hard to identify anyway.