Tutorial Session Information
ISPRS 2026
Tutorial Sessions
Advance Your Expertise with Full-Day and Half-Day Learning Opportunities
Join us on Saturday, 4 July, and Sunday, 5 July, 2026, for a series of tutorial sessions designed to deepen your knowledge and practical skills in both emerging and foundational areas of photogrammetry, remote sensing, and spatial information sciences.
Led by international experts from academia, industry, and government, these full-day and half-day tutorials offer valuable opportunities for hands-on learning, in-depth discussion, and professional networking—just ahead of the main Congress.
Tutorial registration is completed through the Congress registration portal.
Please note that each tutorial session has a limited capacity.
Registration will be on a first-come, first-served basis until all spots are filled.
For any questions or assistance, please contact isprs2026-registration@icsevents.com
Tutorial Information:
Saturday, 4 July, 2026
8:30am - 5:00pm
Tutorial Lead:
Rongjun Qin
Co-Instructors:
Shuang Song
Jiyong Kim
3D Reconstruction from Multi-View Satellite Imagery: From Classic to Modern Methods
Full Day (8 Hours)
High-resolution (high-res) satellite images, which can now observe the Earth with a footprint as small as 30 cm (e.g., WorldView-3/4 sensors), play an important role in generating 3D data over wide areas. Due to orbital limitations, these images have unique characteristics, and their photogrammetric processing requires special considerations. This tutorial is intended for beginners and intermediate users, including students, researchers, and practitioners interested in 3D mapping with multi-stereo satellite images and their applications. Participants will gain hands-on experience with relevant software and learn about recent developments in field-based methods for 3D reconstruction using satellite imagery, including 3D Gaussian Splatting and Neural Radiance Fields (NeRF). The tutorial will cover both theoretical and practical aspects of processing multi-view satellite images and will be divided into a theory session and a hands-on practical session. The theory session will cover: - Existing satellite sensors - Camera models (rational polynomial models) - Bundle adjustment - Surface generation - Orthophoto rectification - Emerging techniques like Neural Radiance Field (NeRF) and 3D Gaussian Splatting (3DGS) In the practical sessions, we will demonstrate: -Satellite imagery selection - Level-1 geometric correction - Pan-sharpening - Relative and absolute orientation - Dense matching using the RPC stereo processor (RSP) to generate colored point clouds, digital surface models (raster and textured meshes), and true-orthophotos. - Examples of NeRF & 3DGS to generate a DSM, shadow mask, and albedo (shadow-free) maps.
Saturday, 4 July, 2026
8:30am - 5:00pm
Tutorial Lead:
Erica Nocerino
Co-Instructors:
Fabio Menna
Dimitrios Skarlatos
Panagiotis Agrafiotis
Gottfried Mandlburger
Caterina Balletti
Katja Richter
Hans-Gerd Maas
Christian Mulsow
A Full Immersion in 3D Underwater Mapping
Full Day (8 Hours)
3D underwater mapping of oceans and inland environments is crucial for a wide range of applications, including marine ecology, archaeology, coastal monitoring, engineering, and hydrology. This tutorial will introduce participants to the latest developments in 3D underwater mapping—both from above and below the water surface—using passive and active sensing techniques. Lectures will cover the theoretical aspects of underwater photogrammetry, through-water optical bathymetry, and underwater image restoration and enhancement approaches. Hands-on practice will engage participants in using the 3D underwater simulation framework POSER (ISPRS ECBI 2024) and sharing benchmark datasets via the NAUTILUS portal (ISPRS SI 2023). Participants will gain practical experience in both photo- and laser-based bathymetry, as well as underwater color correction methods. By providing theoretical and practical insights into simulation, open datasets, and advanced processing and restoration techniques, the tutorial will promote a broad understanding of the challenges in underwater mapping and emphasizes the importance of data validation and sharing. By the end of the tutorial, participants will: · Deepen their understanding of underwater photogrammetry and optical bathymetry principles; · Explore image restoration and color correction methods; · Practice using the POSER simulation framework to design and test 3D underwater mapping scenarios; · Access and utilize benchmark datasets through the NAUTILUS portal for training and evaluation.
Saturday, 4 July, 2026
8:30am - 5:00pm
Tutorial Lead:
Mozhdeh Shahbazi
Co-Instructors:
Victor Al-Hassan
Mikhail Sokolov
Geospatial Deep Learning in Practice
Full Day (8 Hours)
Geospatial Artificial Intelligence (GeoAI) is an emerging field that combines geographic data with machine learning and deep learning techniques to derive insights from spatial information. It is a rapidly growing area driven by advances in remote sensing, big data, and AI technologies. This tutorial will focus on geo deep learning and its application to raster-like data sources, such as satellite imagery, airborne imagery, and elevation models. A short introduction to deep learning will be first given to provide context for participants with little experience in machine learning. This will be followed by a hands-on exercise involving semantic segmentation on large geo-referenced images (GeoTIFF/COG formats) using an open-source geo‑deep‑learning framework developed by Natural Resources Canada. Participants will learn how to use this framework through a complete workflow: · Transforming large high‑resolution imagery into manageable datasets; · Training models with various architectures including convolutional neural networks, transformers, and masked auto-encoders, using scalable solutions for handling large datasets, tracking experiments and optimizing hyperparameters; · Geo-Inference using advanced techniques to optimize performance both in processing time and resource consumption. This tutorial is suitable for all levels. Beginners will gain a solid introduction to geospatial deep learning, while advanced users will explore operationalization, scalability, and performance optimization.
Saturday, 4 July, 2026
8:30am - 5:00pm
Tutorial Lead:
Ewelina Rupnik
Co-Instructors:
Mehdi Daakir
Marc Pierrot-Deseilligny
Hybrid and Precise Camera Pose Estimation in MicMacV2
Full Day (8 Hours)
This tutorial aims to introduce participants to the open-source photogrammetric processing suite MicMacV2 (MMVII) through two datasets that combine hybrid observations (i.e., photogrammetry and surveying) and camera models (perspective camera and pushbroom sensor), along with hands-on programming examples. This tutorial will mark MicMacV2’s first international release. In development since 2021, MicMacV2 is the successor to MicMacV1. This new version is designed to provide an advanced photogrammetric tool that is accessible both to end users (experts and students) and to external developers interested in contributing to the project. The tutorial will include three sessions: · The first session will focus on the metrological aspects of camera pose estimation, such as simultaneous adjustment of photogrammetric and topometric/surveying observations. · The second session will focus on refining the parameters of the aerial perspective camera model and the RPCs of the satellite pushbroom sensor in a unified adjustment. · This final session (in Python or C++) will introduce participants to the MicMacV2 programming environment, including the basic mechanisms for using the library and, optionally, how to add a new command to the tool.
Saturday, 4 July, 2026
8:30am - 5:00pm
Tutorial Lead:
Andreas Piter
Co-Instructors:
Mahmud H. Haghighi
Alison Seidel
InSAR Time Series Analysis with SARvey and InSAR Explorer
Full Day (8 Hours)
InSAR is a powerful tool in engineering, enabling accurate assessment of ground deformations and structural stability. This tutorial offers a hands-on introduction to two open-source tools for analyzing and visualizing InSAR time series: SARvey and InSAR Explorer. SARvey is a software package designed for single-look InSAR time-series analysis, with a focus on detecting and monitoring deformations in engineering applications. Typical use cases include assessing dam stability, monitoring roads and railways, and mapping urban deformations at the building scale. The tutorial will provide a complete SARvey workflow—covering installation, parameter configuration, and advanced processing methods—making it an excellent entry point for newcomers to InSAR, as well as a valuable resource for experienced users seeking more advanced analytical capabilities. InSAR Explorer complements SARvey as a QGIS plugin, allowing smooth integration of InSAR-derived deformation data into Geographic Information Systems. The plugin provides intuitive tools for mapping, overlaying auxiliary datasets, and comparing outputs from various processing workflows. Its user-friendly interface enables quick visualization of deformation time series, creation of interactive plots, and in-depth evaluation of results. In this tutorial, participants will use notebooks in a Google Colab environment to follow the entire workflow—from software installation to the execution of real-world case studies using Sentinel-1 data. Attendees will learn how to adjust processing parameters, interpret deformation time series, and use InSAR Explorer within QGIS for data visualization and analysis. Whether participants are new to InSAR or are experienced practitioners exploring new tools, this tutorial will offer them a comprehensive, practical learning experience to advance their skills in Earth observation and deformation monitoring.
Saturday, 4 July, 2026
8:30am - 5:00pm
Tutorial Lead:
Elena Belcore
Co-Instructors:
Paolo Dabove
Alessandro Frigeri
Darshana Rawal
Open-Source Geospatial Tools for Multisensor Environmental Surveying: Positioning, Photogrammetry, Machine/Deep Learning, and Data Security
Full Day (8 Hours)
This tutorial presents key open-source tools for geospatial education, focusing on GNSS positioning, photogrammetry, AI-based image analysis, and data security. Designed for university-level teaching, it emphasizes accessible solutions for instructors and students. The tutorial will include four modules comprised of lectures and hands-on exercises based on open datasets that will be provided by the instructors. Module 1 - GNSS Positioning : This module will cover the basics of satellite-based geolocation, including positioning accuracy, signal types, and data integration. Participants will use open-source tools, such as RTKLIB, to collect, process, and visualize GNSS data, gaining experience with real-world geospatial workflows. Module 2 - Planetary Mapping: This module will introduce photogrammetric techniques and their application in an open-source software (OpenDroneMap). It will also introduce the principles of planetary geological mapping, with a focus on mapping celestial bodies like the Moon using an open-source software (FOSS). Module 3 - Information Extraction with Deep Learning: This module will explore object detection and classification of UAV imagery using machine and deep learning via several QGIS plugins (Deepness, TreeEyed, SegMAp) and prepared scripts in Jupyter notebooks. Module 4 - Geospatial Data Security: In this module, participants will learn how to protect geospatial data during storage and transfer by integrating blockchain technology with cryptographic techniques for securely encrypting geospatial coordinates. Additionally, they will learn how smart contracts can enforce data-sharing policies and verify user permissions. By the end of the tutorial, participants will learn how to: · Process data using open-source positioning and photogrammetry platforms; · Analyze and interpret data using QGIS and Python; · Use blockchain-based methods for securing geospatial data.
Saturday, 4 July, 2026
8:30am - 5:00pm
Tutorial Lead:
Shahpoor Moradi
Co-Instructors:
Mahkame Moghadam
Sohrab Ganjian
Vittorio Cannas
Stefania Amici
Mozhdeh Shahbazi
Quantum Computing for Earth Observation
Full Day (8 Hours)
Classical high-performance computing has long supported remote sensing and Earth observation activities. However, as data volumes and modeling complexity continue to grow, classical computational approaches are becoming increasingly strained. This trend motivates the exploration of emerging computing paradigms. One promising direction is quantum computing, which leverages quantum mechanical principles such as superposition, entanglement, and interference to perform certain computations more efficiently than classical systems. With this in mind, this tutorial offers a practical introduction to quantum computing, with a focus on quantum machine learning. Participants will have the opportunity to implement a quantum machine learning algorithm for a real-world application in multispectral satellite image analysis. Basic knowledge of linear algebra and Python programming is required to participate in this tutorial. The tutorial will begin with an accessible introduction to the principles of quantum computing. A representative problem based on satellite imagery will then be defined, and the tutorial will alternate between conceptual discussions and hands-on implementation, including gate-based quantum circuit design. As the session progresses, the interplay between quantum and classical resources will be explored across the full processing pipeline. Simulated implementation of a quantum system will be introduced, and with the support of PINQ², access to real quantum platforms of IBM will be made possible as well. By the end of the tutorial, participants will gain a foundational understanding of quantum computing and its practical applications in remote sensing.
Saturday, 4 July, 2026
8:30am - 12:00pm
Tutorial Lead:
Katharina Anders
Co-Instructors:
Bernhard Höfle
Xiaoyu Huang
Ronald Tabernig
Open-source Scientific Software py4dgeo for Change Analysis in 3D/4D Point Clouds
Half Day (4 Hours)
This tutorial introduces py4dgeo, an open-source Python library for analyzing geometric change and surface dynamics in 3D and 4D point cloud data. Designed to support scientific workflows, py4dgeo offers a reproducible and scalable tool for quantifying surface change in multitemporal point clouds and 3D time series across a broad range of topographic monitoring applications. Participants will gain a solid understanding of the key concepts and challenges in 3D/4D change analysis. This includes the full pipeline of multi-temporal point cloud alignment, 3D change quantification, and time series-based quantification methods. The tutorial will demonstrate how py4dgeo implements state-of-the-art algorithms to address these challenges. It will emphasize the library’s modular framework, offering accessible and configurable methods that empower users to perform transparent, fully automated scientific analysis. This tutorial is intended for researchers, students, and practitioners working with time-dependent 3D data who require a flexible, scalable, and open-source framework for surface change analysis. Hands-on exercises will guide participants through practical use of the library, including: · Loading point cloud data, · Applying 3D change detection algorithms (e.g., M3C2), · Using a hierarchical approache for 3D change analysis, · Performing time series-based analysis (e.g., 4D objects-by-change), and · Visualizing results. Example workflows will demonstrate how py4dgeo integrates with standard Python-based environments and complements other open-source tools such as CloudCompare. By the end of the tutorial, participants will understand core methods for 3D/4D change analysis, be able to reproduce and adapt workflows for research or applied monitoring tasks and apply py4dgeo to their own datasets.
Saturday, 4 July, 2026
1:00pm - 4:30pm
Tutorial Lead:
Ayman Habib
Co-Instructors:
Mohamed M.R. Mostafa
Photogrammetric Mapping by Drones: Theory and Practice
Half Day (4 Hours)
This tutorial offers foundational insights into the intricacies of Surveying with Drones. It will address the design, development, integration, operation, and calibration of drone systems, along with best practices for achieving optimal accuracy with various payloads across a range of real-world applications. These include high-definition mapping for autonomous vehicles, civil engineering, mining, digital forestry, precision agriculture, and general surveying and mapping. The intended audience includes students, educators, technicians, engineers, surveying and mapping professionals, and decision-makers. The tutorial will cover the fundamentals of photogrammetry, multi-sensor fusion, and drone positioning and sensor georeferencing. A key focus will be on the integration of multiple imaging and navigation sensors, such as RGB, NIR, and thermal cameras, as well as GNSS and inertial systems. When properly integrated—either during post-mission processing or in near-real-time—these systems can produce high-precision mapping products, enabling diverse information extraction scenarios and delivering valuable insights across multiple application domains. The tutorial will also highlight the influence of technological challenges on day-to-day drone operations in relation to: · Airframe selection, · Mission planning, · Data acquisition, · Data processing, · Calibration, · Quality control, and · Accuracy assessment. Q&A sessions will be encouraged during each section of the tutorial to foster interactive learning and address participants' specific questions.
Sunday, 5 July, 2026
8:30am - 12:00pm
Tutorial Lead:
Ayman Habib
Co-Instructors:
Songlin Fei
Digital Twinning with UAV and Backpack Mobile Mapping Systems
Half Day (4 Hours)
This tutorial provides a comprehensive overview of using photogrammetric and LiDAR sensors integrated onboard Unmanned Aerial Vehicles (UAVs) and wearable backpack systems for generating digital twins across diverse environments, including urban, natural, and mixed landscapes. The tutorial will begin with an in-depth discussion on sensor integration, emphasizing hardware configurations that combine optical imaging and LiDAR systems with GNSS/INS units. It will address the synchronization of cameras, LiDAR units, GNSS, and IMU sensors to ensure precise and reliable data capture. The discussion will proceed to system calibration, focusing on geometric calibration procedures. Participants will learn methods for refining both intrinsic and extrinsic parameters of cameras and LiDAR sensors — steps that are essential for maintaining data integrity and accuracy. The tutorial will continue with georeferencing techniques including: Direct georeferencing using GNSS/INS; Indirect georeferencing with ground control points (GCPs); Simultaneous Localization and Mapping (SLAM); and Trajectory Enhancement and Mapping (TEAM). It will also discuss strategies for integrating UAV and backpack datasets, which often differ in accuracy and spatial reference frames. The tutorial will also cover 3D modeling workflows, including point cloud generation, surface reconstruction, and feature extraction. Particular attention will be given to the challenges and solutions involved in fusing photogrammetric and LiDAR datasets to create high-resolution, geometrically accurate digital twins. Practical examples will demonstrate how multi-scale datasets can be used in applications such as urban infrastructure assessment, forest inventory, and disaster response planning. By the end of the tutorial, participants will gain hands-on knowledge of the complete workflow—from data acquisition to the creation of actionable digital twin models for both research and operational applications.
Sunday, 5 July, 2026
8:30am - 12:00pm
Tutorial Lead:
David Youssefi
Co-Instructors:
Valentine Bellet
Dimitri Lallement
Getting Started with CNES Open-Source 3D Tools in Python
Half Day (4 Hours)
This hands-on tutorial presents a complete open-source workflow for generating and analyzing 3D geospatial data from stereo satellite imagery. It will begin with essential theoretical foundations, including satellite orbits and the principles of photogrammetry, before introducing three key tools: CARS, Bulldozer, and xDEM. This tutorial is designed for a broad audience, including beginners, students, geospatial analysts, researchers, and professionals interested in Earth observation, photogrammetry, and 3D mapping. It will offer an accessible introduction to advanced remote sensing techniques using open-source tools and real-world data. No prior experience is required to attend this tutorial; basic knowledge of Python is helpful but not necessary. Using interactive Python notebooks, participants will: · Generate a Digital Surface Model (DSM) with CARS, creating a 3D representation of the Earth's surface as observed by satellites; · Extract a Digital Terrain Model (DTM) with Bulldozer, isolating the bare ground from the DSM; Participants can also derive a Digital Height Model (DHM) to emphasize above-ground features such as buildings and vegetation — enabling applications like building height estimation for digital twin environments; · Analyze 3D products with xDEM, performing post-processing tasks such as coregistration and detailed change detection. By the end of the tutorial, participants will learn how to produce and interpret DSM, DTM, and DHM products, apply photogrammetric concepts, and perform advanced spatial analyses using satellite-derived 3D data.
Sunday, 5 July, 2026
8:30am - 12:00pm
Tutorial Lead:
Konstantin Klemmer
Co-Instructors:
Marc Rußwurm
Esther Rolf
Evan Shelhamer
Towards Geospatial Embeddings: Investigating Accurate and Accessible Deep Geospatial Feature Representations
Half Day (4 Hours)
This tutorial introduces participants to Earth Embeddings — a new class of deep geospatial representations that unify heterogeneous remote sensing and environmental data into a shared, learnable embedding space indexed by spatiotemporal coordinates. The tutorial is structured into two thematic blocks, each combining concise lectures by leading researchers with practical hands-on sessions. Block 1: Geospatial Embedding Fields This block will cover the foundations of Geospatial Neural Encoding Fields (GeoNEFs), which generate location-specific feature representations from multimodal geospatial data. Participants will learn how embedding fields are constructed using lightweight and foundation vision encoders, and how these embeddings support downstream geospatial tasks. A hands-on session will guide participants through generating embeddings from satellite imagery and environmental time series using pre-trained models. Block 2: Location Encoding and Implicit Neural Representations This block will introduce implicit neural representations and coordinate-based location encoding models that enable scalable geospatial embeddings via neural networks. Participants will explore pre-trained models such as SatCLIP and GeoCLIP, and will learn how to integrate these representations into various downstream use cases. A hands-on tutorial will demonstrate how to build and query Earth Embedding models using open-source libraries and geospatial datasets.
Sunday, 5 July, 2026
8:30am - 12:00pm
Tutorial Lead:
Ruisheng Wang
Co-Instructors:
Fabio Remondino
Liangliang Nan
Gunho Sohn
Florent Larfarge
Urban Scene Modeling
Half Day (4 Hours)
This tutorial offers an in-depth exploration of cutting-edge techniques and applications in 3D urban modeling and digital twin technologies. The objectives of the tutorial include: · Enhancing participants' knowledge of urban modeling methodologies, · Promoting the integration of geospatial data and analytics, · Stimulating innovation in digital twin environments. By showcasing recent advancements, the tutorial aims to support the development of state-of-the-art technologies to address complex urban challenges. Its significance lies in the growing demand for accurate, interactive, and scalable urban models that inform planning, management, and decision-making in smart cities. This tutorial will serve as a platform for interdisciplinary collaboration, encouraging the exchange of ideas among professionals from academia, government, and industry. By the end of the tutorial, participants will gain a solid understanding of the key techniques in 3D point cloud processing using AI as well as data acquisition and processing for urban modeling. Through use-case demonstrations, expert presentations and a panel discussion, attendees will gain valuable insights into the current landscape and future directions of urban scene modeling and digital twin development.