Stefan Laible

About

Stefan Laible is a leading researcher in agricultural robotics and autonomous navigation, specializing in sensor fusion and machine learning for outdoor terrain perception. His work centers on enabling mobile robots to robustly detect, classify, and navigate complex natural environments, with a key application in precision agriculture for crop scouting and weed control. Laible’s major contributions include pioneering the use of 3D LIDAR and camera data fusion for plant species classification—his most cited paper (64 citations) demonstrates a novel method for distinguishing plant species, a critical step toward automated weeding. He further advanced terrain classification by developing conditional random field models and recurrent neural networks that maintain accuracy under varying lighting and surface conditions, as seen in his 39- and 29-citation papers. His research on spatio-temporal classification for semantic robot localization (8 citations) has also improved how robots build local terrain maps for safe, efficient navigation. Collectively, Laible’s work has laid foundational techniques for robust visual terrain recognition, directly impacting the reliability of autonomous agricultural vehicles and outdoor mobile robots.

Research Focus

Key Achievements

Top Papers

1. 1
   Plant Species Classification Using a 3D LIDAR Sensor and Machine Learning

   64 citations · 2010
2. 2
   3D LIDAR- and Camera-Based Terrain Classification Under Different Lighting Conditions

   39 citations · 2012
3. 3
   Terrain classification with conditional random fields on fused 3D LIDAR and camera data

   29 citations · 2013
4. 4
   Robust Visual Terrain Classification with Recurrent Neural Networks

   9 citations · 2015
5. 5
   Building local terrain maps using spatio-temporal classification for semantic robot localization

   8 citations · 2014

Key Collaborators

Contact & Links