Invited Talks


Prof. Rafael Garcia
University of Girona

Vision-based Robots to Explore the Ocean

The ocean is covering nearly 75% of our planet. However, humans have explored a very small fraction of this vast extension of the globe, since the deep sea is a hostile environment with absence of light and extreme pressures. The characteristics of the underwater environment offer several challenges for computer vision, mainly due to the significant attenuation and scattering of visible light. Commonly, underwater images suffer from lack of image contrast, blurring or strong moving shadows from artificial illumination. Moreover, light attenuation produces images with limited range. But even if using vision underwater is a difficult endeavor due to the transmission properties of the medium, with the adequate processing pipeline, vision can be a powerful tool for underwater robots to explore the ocean.
In this talk we will see how image enhancement techniques can be applied to successfully restore underwater images, and how to overcome the different problems associated to the interaction between light and water to successfully explore and map unknown areas of the seafloor.


Dr. Larry Matthies

Progress Toward Autonomous Aerial Mobility Applications in Planetary Exploration

Aerial mobility has been under study for planetary exploration since at least 1970. The Soviet Union flew balloon missions in the Venus atmosphere in 1985. Since then, balloon missions have been studied for Mars and Titan, fixed wing aircraft missions have been studied for Mars, Venus, and Titan, and rotorcraft have been studied for Mars and Titan. Recent progress on autonomous navigation of small rotorcraft (order of 1 kg) on Earth has inspired new ideas for small rotorcraft as scouts for Mars rovers and as daughtercraft to deploy from landers or balloons on Titan, as well as for guided gliders to deploy from balloons on Venus. This progress is partly enabled by the rapid progress in miniaturization and increased performance of microprocessors and sensors for mobile consumer electronics. This talk will give an introduction to concepts for such missions, as well as show progress toward relevant capabilities that is being made under micro air vehicle projects for both Earth and space applications, including capabilities for visual-inertial navigation, vision-based autonomous safe landing, and vision-based obstacle avoidance.


Prof. Tomas Pajdla
CTU Prague

3D Reconstruction on Planets

We will review image data available form some current as well as historical missions on Mars, Moon and other planets, present some results of 3D reconstructions, and discuss open challenges for automatic 3D reconstruction from planetary data.


Prof. Marc Pollefeys
ETH Zurich

Vehicle localization and mapping with monocular wide-field-of-view cameras


Dr. Andreas Wendel
Google [X]

Realizing Self-Driving Cars

© 2015 Workshop on Computer Vision in Vehicle Technology