Background

Kovilta Oy was founded in 2011 by D.Sc. Jonne Poikonen, D.Sc. Mika Laiho and Prof., D.Sc. Ari Paasio. The team has a long research background from the University of Turku and Helsinki University of technology (Aalto University), in the fields of integrated circuit design, parallel visual analysis systems and image processing algorithms.

The Kovilta team has extensive experience in the design, implementation and testing of full-custom ASICs, efficient methods and parallel algorithms for image analysis, FPGA- and embedded system design and software design. Through continuous involvement in state-of-the-art research, and an open-minded, "can-do" attitude, we are able to efficiently recognize and utilize new enabling technologies not yet widely applied by large machine vision manufacturers.

Kovilta's technology development has been funded in part by TEKES, the Foundation for Finnish Inventions (Keksintösäätiö), and the Runar Bäckström Foundation.

Contact:

Jonne Poikonen (CEO)

Kovilta Oy
Piispanristintie 1
FI-20760 Piispanristi
Finland

Tel: +358-505905490
Email: firstname.lastname@kovilta.fi

VAT-ID: FI23821450




Research and development

An essential part of Kovilta's corporate philosophy is a continous strong emphasis on new technological and application-level development. To this end Kovilta takes part in relevant research and development projects in collaboration with other industrial or academic partners, in addition to the Kovilta's in-house development.

Kovilta took part in the STREAKDET project, funded by the European Space Agency (ESA) and lead by the Finnish Geodetic institute. Kovilta's responsibility in the project was the development of image segmentation methods and algorithms for space debris detection from telescope data.

Kovilta has also been involved as an industrial partner in two Tekes-funded research projects lead by University of Turku: MARIN and AgiSpacES. The MARIN project focused on Augmented Reality (AR) technology, where Kovilta's single-chip, low-power processor platform could provide a more efficient tool for extracting visual correspondence points between real-life and simulated data.