Vision for Robotics Lab

Open Opportunities

Label transfer from satellite to aerial imagery ETH Zurich Autonomous Systems Lab Other organizations: Vision for Robotics Lab The project aims to implement a semantic label transfer from satellite to aerial imagery in order to enable the training of image-based machine learning algorithms for autonomous aerial vehicle tasks, such as path planning, collision avoidance, and localization. Computer Vision, Intelligent Robotics Semester Project

Probabilistic Air-ground Localization in Agricultural Environment ETH Zurich Vision for Robotics Lab Digital environments, or digital twins, allow for design, prototyping, and testing in the virtual world before moving to the real world, thus accelerating development and reducing costs. Digital twin of a farm supports crop operations such as scheduling a harvest or predicting a yield, while agritech companies can develop farm automation robots using a digital twin. The goal of this project is to develop air-ground localization strategies that are capable to be deployed in crop environments during the production season. The main target is to identify individual plants in the ground images using as reference the aerial images. Agricultural Engineering, Computer Vision, Intelligent Robotics Master Thesis, Semester Project

Semantic Segmentation-Aided Bundle Adjustment ETH Zurich Autonomous Systems Lab Other organizations: Vision for Robotics Lab Bundle Adjustment (BA) is a critical optimization technique used to refine a visual reconstruction by jointly estimating the 3D scene structure and the viewing parameters. Traditional BA approaches primarily focus on geometric features and might struggle in highly unstructured scenarios, such as natural environments. This project aims to extend the Bundle Adjustment methodology by incorporating higher-level features extracted from semantic segmentation. The integration of semantic information aims to provide contextually relevant and more discriminative data to the adjustment process, thereby improving its accuracy and robustness. Computer Vision, Image Processing Master Thesis, Semester Project

Domain-Adaptive Semantic Segmentation for Robotic Navigation in Natural Environments ETH Zurich Autonomous Systems Lab Other organizations: Vision for Robotics Lab Natural environments are inherently dynamic and heterogeneous, characterized by significant variability in appearance and structure. The primary goal of this project is to develop a training scheme capable of adapting lightweight semantic segmentation models to changing environments without relying on manually annotated data. In particular, the student will investigate advanced domain adaptation techniques to bridge the gap between synthetic and real data, as well as between sets of real images collected during different seasons. The developed approach is expected to enhance the model’s generalization capabilities, allowing it to adapt more rapidly to unseen conditions, and ultimately improving the robots' operational efficacy and robustness. Computer Vision, Image Processing Master Thesis, Semester Project

3D Reconstruction of Natural Environments for Robotic Navigation ETH Zurich Autonomous Systems Lab Other organizations: Vision for Robotics Lab In this project, the student will explore efficient ways of modeling a natural environment to support long-term robotic navigation. Computer Vision, Intelligent Robotics Master Thesis, Semester Project