Vision2Move

Digitization is permeating core areas of both businesses and municipalities. With the goal to create smart cities and regions, multimedia sensor systems are being installed on a large scale. By evaluating the collected data using machine learning, particularly computer vision as in the Vision2Move project, innovative applications in the mobility sector can be developed. By processing raw data, comparable to the refining of petroleum, both new opportunities and new challenges arise in the mobility domain. Particularly promising is the interlacing of different modalities of data, which is precisely the area addressed by the Vision2Move JR Center.
 

Scientific Challenges
 

The merging and interlacing of individual observations of objects (e.g., vehicles, animals, people) by several sensor systems is not yet possible with satisfactory accuracy. As a result, it is not yet feasible to use machine-based traffic counting or traffic control systems at the application level.

• For certain use cases (e.g., obstacle detection on railroad lines), there is not enough corresponding data to train the algorithms intended for this purpose. Therefore, it is necessary to investigate how synthetic data can be generated for the purpose of providing computer vision algorithms with sufficiently large training data.
• Another challenge is the efficient and effective processing of data on the basis of a limited number of visual sensor units, for example to comply with the GDPR, but also to provide for energy-efficient operation of systems.

Areas of Application

Research on these challenges will contribute to the following areas of application:

• Traffic monitoring and guidance at the local and regional level
  Wildlife detection and monitoring
  Establishing safe traffic areas for automated mobility
• Recognition of behavioral patterns and their assessment
  Detection of near misses
  Detection of unsafe driving behavior
  Detection of abnormal behavior of animal herds (such as flocks of sheep during a wolf attack)
  Support in the search for persons suffering from dementia