3D Remote Sensing Reconstruction and Visualisation Training Course

Participants can establish spatial reference and control arrangements for accurate 3D products.

• Horizontal and vertical datums selected for project location and purpose
• Coordinate reference system metadata checked before importing source datasets
• Ground control points designed for distribution, visibility and stability
• Independent checkpoints separated from control points for accuracy testing
• GNSS observation quality, antenna height and epoch recorded systematically
• Geoid models applied when converting ellipsoidal heights to orthometric elevations

Participants can select suitable sensors and acquisition parameters for reconstruction objectives.

• RGB, multispectral, thermal and LiDAR sensors matched to information needs
• Ground sample distance calculated from focal length, altitude and pixel size
• Forward and side overlap set for terrain, buildings and vegetation
• Flight-line geometry planned to minimise occlusions on complex structures
• LiDAR pulse density related to canopy penetration and feature detection
• Sun angle, wind, cloud and seasonal conditions assessed before capture

Participants can inspect, organise and prepare imagery and LiDAR files for processing.

• Image EXIF metadata reviewed for focal length, position and timestamps
• Camera calibration information assessed before photogrammetric processing begins
• LAS and LAZ headers inspected for coordinates, classification and point counts
• File naming conventions linked to acquisition blocks and processing versions
• Blurred, overexposed and obstructed images identified through systematic screening
• Storage architecture separates raw, intermediate, validated and published datasets

Participants can explain and diagnose the geometric principles behind image-based reconstruction.

• Central projection relates image coordinates to three-dimensional object space
• Tie points established through feature detection and image matching
• Bundle adjustment estimates camera positions, orientations and calibration parameters
• Camera interior orientation distinguished from exterior orientation parameters
• Parallax interpreted when estimating depth from overlapping stereo imagery
• Weak geometry recognised in linear corridors, uniform surfaces and nadir-only imagery

Participants can produce and evaluate a sparse reconstruction from overlapping imagery.

• Feature detectors selected according to texture, scale and image quality
• Image matching thresholds adjusted to reduce false correspondences
• Sparse point clouds inspected for gaps, noise and geometric distortion
• Reprojection error used to identify problematic cameras and tie points
• Ground control observations introduced after initial image alignment
• Camera alignment reports documented with residuals and rejected observations

Participants can generate dense point clouds and select settings appropriate to intended use.

• Dense image matching converts aligned imagery into three-dimensional point clouds
• Depth-map resolution balanced against processing time and required detail
• Noise filtering removes isolated points without erasing valid small features
• Point density compared against ground sample distance and surface complexity
• Occlusions and shadowed surfaces identified within dense reconstruction outputs
• Point-cloud exports prepared using LAS, LAZ and E57 formats

Participants can prepare LiDAR point clouds for surface modelling and feature extraction.

• Discrete-return and full-waveform LiDAR characteristics compared for target detection
• Point-cloud registration aligns overlapping terrestrial or mobile scanning stations
• Strip adjustment identifies systematic elevation differences between flight lines
• LAS classification codes interpreted for ground, vegetation and built features
• Outlier detection removes isolated returns and sensor-related artefacts
• Intensity values assessed as supplementary evidence rather than surface colour

Participants can classify and segment point clouds using defensible rules and review procedures.

• Progressive TIN densification separates ground points from non-ground returns
• Cloth simulation filtering compared with slope-based ground classification
• Classification thresholds adjusted for urban terrain, forests and embankments
• Building, vegetation and utility candidates segmented using geometric characteristics
• Manual review targets breaklines, retaining walls and complex roof forms
• Class codes validated against ASPRS LAS classification conventions

Participants can derive terrain and surface models suited to analytical and visual purposes.

• Digital terrain models interpolated from classified ground-point observations
• Digital surface models retain buildings, vegetation and above-ground structures
• Triangulated irregular networks preserve abrupt terrain discontinuities and breaklines
• Raster cell size selected from point spacing and required feature scale
• Hillshade, slope and aspect derived with documented azimuth parameters
• Hydrologically conditioned terrain surfaces prepared for drainage and flood analysis

Participants can produce orthorectified imagery and assess its positional and radiometric quality.

• Orthorectification removes relief displacement using camera models and elevation surfaces
• Seamline placement reduces visible joins across buildings and moving objects
• Radiometric balancing corrects exposure differences between adjacent source images
• True orthomosaics distinguished from conventional orthophotos over vertical structures
• Positional checks compare identifiable features against independent control observations
• Cloud Optimized GeoTIFF exports prepared with suitable pyramids and compression

Participants can create textured 3D meshes while managing geometric complexity and visual quality.

• Poisson reconstruction and Delaunay triangulation compared for mesh generation
• Mesh decimation reduces polygon counts while preserving critical geometry
• Holes, non-manifold edges and self-intersections identified during mesh inspection
• Texture atlases built from aligned imagery and corrected for seams
• UV mapping links image textures to mesh surface coordinates
• OBJ, PLY and glTF exports selected for intended delivery platforms

Participants can assess 3D reconstruction quality and report accuracy against stated requirements.

• Root mean square error calculated from independent checkpoint residuals
• Horizontal, vertical and three-dimensional accuracy reported separately where appropriate
• Systematic bias distinguished from random error through residual pattern review
• Completeness checks identify missing surfaces, occluded zones and data gaps
• Visual inspection criteria documented for meshes, orthomosaics and terrain models
• Quality-control reports link acceptance criteria, evidence and corrective actions

Participants can compare repeat surveys and distinguish significant change from processing artefacts.

• Co-registration aligns repeat datasets before calculating surface differences
• Cloud-to-cloud distance measures quantify local three-dimensional displacement
• DEM of Difference analysis identifies erosion, deposition and construction change
• Minimum level of detection incorporates uncertainty from both survey epochs
• Temporal point-cloud classification separates genuine change from vegetation variation
• Change maps use thresholds, confidence statements and interpretable colour scales

Participants can design 3D visualisations that support technical interpretation and decision-making.

• Viewpoints selected to reveal slope, clearance, extent and spatial relationships
• Vertical exaggeration applied cautiously and disclosed within terrain visualisations
• Colour ramps chosen for elevation, intensity, classification and analytical meaning
• Clipping planes expose subsurface, interior and corridor features clearly
• Level of detail matched to audience, device capability and decision task
• Annotations distinguish measured evidence from interpretive or proposed features

Participants can prepare 3D data for efficient publishing and controlled organisational access.

• 3D Tiles structures stream large meshes and point clouds efficiently
• Scene layers organised by asset type, scale and access requirements
• Coordinate transformations verified before publishing web-based three-dimensional scenes
• Attribute links connect visual features with inspection and asset records
• Caching and level-of-detail strategies improve rendering on limited bandwidth
• Access controls protect sensitive sites, critical infrastructure and personal information

Participants can document 3D datasets for reuse, audit and procurement acceptance.

• ISO 19115 metadata records lineage, quality, reference systems and constraints
• ISO 19157 quality elements applied to positional accuracy and completeness
• Processing logs capture software versions, parameters and source-data identifiers
• Data dictionaries define classes, attributes, units and allowable values
• Retention schedules distinguish raw observations from derived delivery products
• Licensing conditions reviewed for imagery, basemaps and third-party point clouds

Participants can write and evaluate technical requirements for outsourced 3D reconstruction work.

• Statement of work defines coverage, sensors, control and acceptance deliverables
• Accuracy requirements expressed using checkpoints, confidence levels and reporting methods
• File formats, coordinate systems and naming conventions specified contractually
• Pilot-area acceptance tests reduce risk before full production mobilisation
• Supplier submissions reviewed against acquisition plans and quality-control evidence
• Handover packages include raw data, processing projects and final outputs

Participants can establish a repeatable operational workflow for 3D geospatial production.

• Workflow diagrams assign responsibilities from acquisition through publication and archiving
• Processing queues prioritise urgent inspection, safety and decision-support requests
• Compute requirements estimated from image count, resolution and reconstruction settings
• Version control separates approved products from exploratory processing iterations
• Exception procedures address missing control, poor imagery and failed alignment
• Performance indicators track turnaround time, rework rates and acceptance outcomes

Participants can build, review and present a complete 3D reconstruction project with validated outputs and documented limitations.

• Project brief translated into reconstruction specifications, control strategy and deliverables
• Source imagery or point clouds processed into validated three-dimensional datasets
• Terrain, mesh, orthomosaic and visualisation products selected for stated decisions
• Accuracy evidence compiled from checkpoints, residuals and visual quality checks
• Metadata and handover documentation assembled for reuse and audit
• Technical presentation communicates findings, limitations, risks and recommended next actions