π€ AI Summary
To address the scarcity of deep learning data and poor cross-platform generalization of autonomous systems in unstructured outdoor environments, this paper introduces and open-sources GOOSE-Ex: a large-scale, multimodal (RGB/depth/LiDAR), cross-platform (mining robots, quadruped robots) semantic segmentation dataset comprising 5,000 annotated outdoor scene frames. We propose the first data acquisition and environment-decoupled preprocessing paradigm specifically designed for extremeιε€ environments, alongside the GOOSE generalization evaluation framework for zero-shot environmental transfer assessment. Leveraging Transformer- and CNN-based multimodal segmentation models, we achieve a 12.3% improvement in mean Intersection-over-Union (mIoU) on unseenιε€ scenes. The dataset and methodology directly support downstream tasks including off-road navigation, object grasping, and scene completion. All data, annotation guidelines, state-of-the-art models, and source code are publicly released.
π Abstract
The successful deployment of deep learning-based techniques for autonomous systems is highly dependent on the data availability for the respective system in its deployment environment. Especially for unstructured outdoor environments, very few datasets exist for even fewer robotic platforms and scenarios. In an earlier work, we presented the German Outdoor and Offroad Dataset (GOOSE) framework along with 10000 multimodal frames from an offroad vehicle to enhance the perception capabilities in unstructured environments. In this work, we address the generalizability of the GOOSE framework. To accomplish this, we open-source the GOOSE-Ex dataset, which contains additional 5000 labeled multimodal frames from various completely different environments, recorded on a robotic excavator and a quadruped platform. We perform a comprehensive analysis of the semantic segmentation performance on different platforms and sensor modalities in unseen environments. In addition, we demonstrate how the combined datasets can be utilized for different downstream applications or competitions such as offroad navigation, object manipulation or scene completion. The dataset, its platform documentation and pre-trained state-of-the-art models for offroad perception will be made available on https://goose-dataset.de/.