Defence in the field of mechanical engineering M. Sc. (Tech) Risto Ojala

Opponent Professor Laura Ruotsalainen, Helsinki University

Custos Professori Kari Tammi, Aalto University, School of Engineering

Contact information  risto.j.ojala@aalto.fi  +358 440 230 445

In the near future, road traffic will experience notable changes due to the adoption of intelligent transportation systems. New technologies, such as connected vehicle solutions, advanced driver assistance systems, as well as automated vehicle technology are developing rapidly. These technologies offer possibilities for improving the safety and efficiency of road traffic around the globe. However, many intelligent transportation applications depend on comprehensive and real-time data of the surrounding road users. This thesis explored methodologies for detecting and localising road users with roadside cameras. Localisation information provided by roadside cameras could be utilised by for example sharing this information with surrounding connected vehicles. This thesis produced concrete solutions for detecting and localising road users from roadside camera views. The presented methodologies have been designed to solve practical issues related to the adoption of the technology. The developed road user detection algorithm was shown capable of operating on inexpensive hardware, unlike the compared algorithms. Another key result was the excellent generalisability of the proposed calibration approach, which enables camera-based localisation. Additionally, the sources of error in camera-based localisation were experimentally analysed. The results of this thesis can be applied in the design and implementation of roadside camera-based computer vision solutions in intelligent transportation systems. The presented methodologies also support the overall development of intelligent transportation applications. Furthermore, the published results provide a solid foundation for further research on the topic