Abstract
Labeled Mobile Laser Scanning (MLS) data are increasingly in demand for applications such as urban planning and autonomous driving. However, there is a severe shortage of manually labeled, high-fidelity MLS datasets, especially from developing countries like India. Deep learning models trained on less complex environments often fail to generalize to Indian scenarios, hindering the development of robust applications. To address this gap, we introduce Ke-MLS, the largest publicly available, expert-labeled MLS LiDAR dataset from Kerala, India. The dataset comprises 48 classes and supports the training of a variety of deep learning models for multiple applications. We evaluate the performance of state-of-the-art 3D deep learning algorithms on Ke-MLS, achieving mean Intersection-Over-Union (mIoU) scores in the range of 70–80%. The dataset is publicly available at https://lidaverse.com/.
Get full access to this article
View all access options for this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
