Image: Object detection, girl walk with a bike, frame over girl with text: 99%

Innovation and high-tech combined

Our research department is working to combine existing ultrasonic technology with our self-developed sensor technology.

Thanks to modern image processing algorithms, supported amongst other things by Deep Learning, obstacles such as potholes or steps going downwards should also be detected in the future and there should be a distinction between surmountable and insurmountable obstacles. The optional relaying of alternative routes is also planned.

Beyond the mere warning function, the included object detection can also actively provide guidance e.g. to a pedestrian crossing, a refuse bin or steps. The focus here is also on everyday suitability, so there will be numerous possibilities to adapt the system to individual needs.

This project is developed together with Graz University of Technology, Institute for Computer Graphics and Vision.

Images for object detection and accessible area

Image: Object detection
Image: Object detection, steps with warning line
Image: Object detection, bus stop and bicycles marked with warning line

In addition, a common problem in cities can be solved. Zebra crossings often lead diagonally across the road rather than at a 90 degree angle, particularly on wide roads, hence blind people run the risk of straying from the zebra crossing. Camera-based detection here can also warn against leaving the side of the zebra crossing. Further information will be published here on an ongoing basis.

Grafik Projekt

Projekt: Streetview for blind people

A system is being developed that enables a user’s exact position to be determined using image data as part of the research project ‘StreetView for visually impaired people’, which was advertised and funded by the Ministry of Transport in the ‘Mobilität der Zukunft’ programme (Mobility of the future) and is handled by the Austrian Research Promotion Agency.

For this purpose, a previously defined area, e.g. a public square or street, is photographed and a 3D map is then generated. Once this is generated, a photo, e.g. automated by a camera-based InnoMake in the future or via a Smartphone App, can be sent as a request to the system and the exact position can be determined using complex “feature matching”.

The project is jointly implemented by the following contributors:

  • Netwiss OG
  • Tec-Innovation
  • Graz University of Technology, Institute for Computer Graphics and Vision
  • Austrian Association in Support of the Blind and Visually Impaired
  • LDRR Apps and Design
Image: 3D model of Tummelplatz in Graz

A 3D model of Tummelplatz in Graz has been created from many individual images using photogrammetry.

Newsletter

Leave us your mail adress for the newsletter