Welcome to my blog
A new home for my notes on 3D vision, LiDAR, and geospatial deep learning, migrating from Notion.
About
Ph.D. research scholar in the GeoAI4Cities Lab, Data Science and Engineering at IISER Bhopal, advised by Dr. Vaibhav Kumar.
I develop deep learning methods for 3D point clouds, LiDAR, and multi-source geospatial data to better understand and plan urban environments.
I am a Ph.D. research scholar in the GeoAI4Cities Lab at Indian Institute of Science Education and Research, Bhopal, advised by Dr. Vaibhav Kumar. My work sits at the intersection of 3D computer vision and geospatial science.
My doctoral research, "LiDAR Point Cloud Perception for Emergency Vehicle Accessibility and Quality-of-Life Mapping in Urban Environments", develops deep learning models for semantic segmentation of mobile LiDAR point clouds, domain adaptation across urban scenes, and the fusion of LiDAR, street-view imagery, and other 2D and 3D geospatial data for applications such as pedestrian-oriented route planning and urban quality-of-life assessment.
Research Interests
Urban Perception Modeling2D & 3D Computer VisionLiDAR Point Cloud ProcessingDeep Learning for Remote SensingSemantic SegmentationLarge Language ModelsUrban ComputingGeospatial AI
News
- New journal paper in Applied Intelligence on deep-learning semantic segmentation of subsampled LiDAR point clouds.
- Our pedestrian-oriented route planning work (IJAEOG 2025) was featured by Dainik Jagran, covering the data-driven framework leveraging deep learning and spatial perception for safer, scenic navigation routes.
- Founded Netrica.AI (Netrica Pvt. Ltd.), an applied-AI venture for urban planning, incubated at IISER Bhopal and IISER Pune.
- Ke-MLS, a large-scale Indian urban mobile-LiDAR dataset, published in Environment and Planning B.
- Three journal papers published across the Int. J. of Applied Earth Observation & Geoinformation and Neurocomputing.
- Won First Prize at the IEEE-GRSS Student Hackathon and reached the Top 5 at the DST-SOI-Microsoft Geospatial Hackathon.
Publications
All paper PDFsJournal Articles
Appl. Intell. · 2026Performance analysis of subsampled LiDAR point clouds using deep learning based semantic segmentationPyare Lal Chauhan, Aakash Singh Bais, Vaibhav Kumar
Applied Intelligence, vol. 56, Article 273
Benchmarks how point-cloud subsampling and compression strategies trade off semantic-segmentation accuracy against compute cost for airborne and mobile LiDAR.
Env. Plan. B · 2026Ke-MLS: A large-scale labeled mobile LiDAR data set from Indian urban regionVaibhav Kumar, Bharat Lohani, Pyare Lal, Aakash Singh Bais, Aditya
Environment and Planning B: Urban Analytics and City Science
Introduces Ke-MLS, a large-scale labeled mobile-LiDAR dataset of Indian urban streetscapes for point-cloud segmentation research.
IJAEOG · 2025A data-driven framework for pedestrian-oriented route planning leveraging deep learning and spatial perceptionPyare Lal Chauhan, Tanishq Kumar Baswal, Vaibhav Kumar
International Journal of Applied Earth Observation and Geoinformation, vol. 144, Article 104932
Plans pedestrian routes optimized for human perception (safe, lively, beautiful) by combining street-view imagery with deep learning and spatial data.
IJAEOG · 2025Deep learning and multi-source 2D and 3D geospatial data for urban quality of life assessmentAyush Dabra, Pyare Lal Chauhan, Vaibhav Kumar
International Journal of Applied Earth Observation and Geoinformation, vol. 144, Article 104838
Fuses multi-source 2D and 3D geospatial data with deep learning to map and assess urban quality of life.
Neurocomputing · 2025Urban multi-domain mixing (UMDMix) based unsupervised domain adaptation for LiDAR semantic segmentationAnurag Nihal*, Pyare Lal*, Vaibhav Kumar
Neurocomputing, Article 131526
Proposes UMDMix, an unsupervised domain-adaptation method that mixes urban LiDAR domains to transfer semantic segmentation across cities.
Conference Papers
IEEE InGARSS · 2023Addressing class imbalance challenge in semantic segmentation of ALS data through performance analysis of RandLA-Net and PointNet++Pyare Lal Chauhan, J. Vijaywargiya, A. M. Ramiya
IEEE India Geoscience and Remote Sensing Symposium (InGARSS)
Examines class imbalance in airborne-LiDAR (ALS) semantic segmentation, comparing RandLA-Net and PointNet++.
* denotes equal contribution.
Open Source & Datasets
Ke-MLS
Co-creator of a large-scale labeled mobile LiDAR dataset for the Indian urban region.
Street-View Imagery Human Perception
Code for street-view imagery human-perception modeling and pedestrian-oriented route planning (svi_perception).
LiDAR Subsampling Benchmark
Benchmark code for subsampling strategies in LiDAR semantic segmentation (LiDAR-Subsampling-Benchmark).
Urban Multi-Domain Mixing (UMDMix)
Code for unsupervised domain adaptation in LiDAR semantic segmentation via urban multi-domain mixing (umdmix-uda).
Reasoning from Scratch
A from-scratch implementation exploring reasoning in large language models (reasoning-from-scratch).
Experience
Doctoral Researcher
Aug 2022 - PresentGeoAI4Cities Lab, IISER Bhopal · Advisor: Dr. Vaibhav Kumar
- Deep learning for LiDAR semantic segmentation in urban environments; emergency-vehicle accessibility mapping using 3D geospatial data.
- Cross-city perception modeling (Mumbai-Paris) and multi-source 2D/3D fusion for urban quality-of-life with a Global-South corrective framing.
- Unsupervised domain adaptation for cross-domain LiDAR semantic segmentation.
M.Tech. Research
2020 - 2022IIST Thiruvananthapuram · Advisor: Dr. A. M. Ramiya
- Semantic segmentation of airborne LiDAR with deep learning (RandLA-Net, PointNet++); addressed class-imbalance in ALS classification.
Research Intern
Jun 2022 - Aug 2022Indian Institute of Soil & Water Conservation (IISWC), Dehradun
- Land-use / land-cover classification and change detection on Sentinel-2 and Landsat imagery; NDVI/EVI time-series workflows on Google Earth Engine.
Data Science Intern
Feb 2022 - May 2022- Road-network detection and pre-fire risk prediction using Sentinel-1 (SAR) and Sentinel-2 (optical) imagery; vegetation-cover-change pipelines in Python.
Teaching
Teaching Assistant, IISER Bhopal: Artificial Intelligence (DSE313), Spatial Data Science (DSE416), Accelerated Applied AI, Applied Optimization.
Venture
Netrica.AI · Netrica Pvt. Ltd.
Feb 2026 - PresentFounder
An early-stage AI venture building applied-intelligence solutions for urban planning, incubated at IISER Bhopal and IISER Pune.
Incubated at the IISER Bhopal (IICE) and IISER Pune (AIC) incubation centres.
Education
Ph.D. in Data Science and Engineering
2022 - PresentIndian Institute of Science Education and Research, Bhopal, Madhya Pradesh
CGPA 9.0
M.Tech. in Geoinformatics
2020 - 2022Indian Institute of Space Science and Technology, Thiruvananthapuram, Kerala
CGPA 8.69
B.Tech. in Civil Engineering
2012 - 2016SRM University, Chennai, Tamil Nadu
CGPA 9.17
Technical Skills
- Programming
- Python, MATLAB
- Deep Learning
- PyTorch, TensorFlow, Point Cloud Networks
- Geospatial
- Google Earth Engine, QGIS, ArcGIS, CloudCompare, Open3D
- Remote Sensing
- Sentinel-2, Landsat, Sentinel-1 SAR, LiDAR (ALS/MLS) processing
- Tools
- Git, Docker, LaTeX, Linux
- Languages
- Hindi (native), English (professional)
Awards & Honors
- 2023
First Prize, IEEE-GRSS Student Hackathon
Winner at the IEEE GRSS Student Hackathon, Bangalore.
- 2023
Top 5, DST-SOI-Microsoft Geospatial Research Hackathon
Top-5 team in the Geospatial Research Hackathon, IIIT Hyderabad.
- 2023
ACM India Anveshan Setu Fellowship
Selected for the ACM India Anveshan Setu Fellowship.
- 2022
Travel Grant, IndoML Symposium
Travel grant for the IndoML Symposium, IIT Gandhinagar.
Blog
Get in Touch
I'm always open to discussing research, collaborations, and new opportunities. Feel free to reach out.
Bhopal, Madhya Pradesh, India