3D Photogrammetry PDF: Introduction, Techniques, and Applications

3D Photogrammetry Introduction to 3D Photogrammetry Photogrammetry + Remote sensing = measuring 3D world utilizing sensors Key tasks: 3D data acquisition, object detection, geometric modeling, semantic/topological recovery. Point Clouds: Sets of discrete 3D points, measuring surfac...

3D Photogrammetry Introduction to 3D Photogrammetry Photogrammetry + Remote sensing = measuring 3D world utilizing sensors Key tasks: 3D data acquisition, object detection, geometric modeling, semantic/topological recovery. Point Clouds: Sets of discrete 3D points, measuring surfaces, with 3D coordinates (each point) & attributes, may contain noise. Point cloud processing uses signal processing, pattern recognition, and machine learning->digital twin Challenges and Key Techniques Challenges: Noise, outliers, complex backgrounds, incomplete objects, uneven densities, lack of topology. Key techniques-processing: o Registration: Aligning multiple datasets, comparion o Segmentation: separation into segments with shared labels/characteristics, extraction o Classification: Assigning labels to points/segments, recognition Acquisition of Point Clouds Point clouds: Geometric (position) and radiometric (attributes) data, consider sensors Recorded as: List of points [x, y, z, attributes](rgb, intensity, number of returns, incidence angle). Active methods(laser scanning): direct, Emit signals (LiDAR), measure distance/angles, calculate 3D positions. Passive methods: Use image pairs or single images. o Image pairs(stereo vision): Use corresponding points for depth via triangulation. o Single images(monocular vision): Use prior knowledge/deep learning. SfM/MVS: Structure from Motion (image pairs), Multi-view stereo (multiple images). Representation Point clouds represent 3D values: elevation, color, intensity. Interpretation needs: geometry, topology, semantics. Segmentation: Partitions point clouds into like labeled segments (objects). Point cloud Definitions Segmentation: partition into segments w same characteristics, each segment unique label, Semantic Segmentation: Partitions into segments with the same characteristics and assigns predefined labels, same labels. Classification: Assigns single label to entire point cloud. Object detection: detects and locates objects of interest. Semantic Understanding with AI AI: Mimics human intelligence. ML: Algorithms learn from data to make predictions, subset of IA DL: Subset of ML using multiple layers for data representation. Supervised learning: Uses labeled data. Unsupervised learning: Does not use labeled data. Clustering – DBSCAN-unsupervised learing DBSCAN: Density-based clustering, using min points & distance (epsilon). Identifies core(close to others)/non-core(away) points, iterativelly. Cluster(segment) if core points are connected Advantages: Handles any shape, no predefined cluster number, outlier filtering, good for spatial data. Disadvantages: Parameter selection non trivila, computationally heavy, memory intensive. Machine Learning Approaches Hand-crafted features: Designed by experts for specific tasks; may not generalize well. Random Forest (RF): Builds decision trees using diff subset data,&features multiple trees-randomization, random subset selection prediction-unseen data passes through tree->majority vote(classification) or average (regression). geometric features: o attributes calculated from the 3D coordinates of points within a point cloud, spatial rel. o higher-level understanding local surf characteristics-> segmentation classification o input data that the decision trees in the RF algorithm use to make predictions o Eigenvalue based: Linearity, Planarity, Scattering o Gom 3D prop: radus of k-nearest neighbour, local point density Hand-crafted features can outperform DL for user specific tasks when data knowledge is available. Deep Learning Approaches Deep learning: Learns features automatically using multiple layers of encoding features. 3D deep learning methods-semantic segmentation models: o View-based: Uses 2D snapshots, CNNs (convolutional nn); scalability issues. o  o Voxel-based: Voxelizes point clouds-enforce struct; computationally heavy. Voxel-3D volum. pixel o Graph-based: Operates on point cloud topologies; computationally heavy. o Point-based: Operates directly on unstructured point sets. PointNet: Point-based, uses T-Nets, MLPs, max pooling; inputs n points, outputs classification scores or semantic segmentation. Data Science in Earth Observation Core Theme: AI's Growing Role in Earth Observation (EO) Explorative signal processing method, date fusion, info mining, ML, DL, big data analytics HPC Key Remote Sensing Technologies Satellite Remote sensing (RS) -enables large-scale, contact-free Earth information gathering. Electromagnetic Radiation (EMR) carries the information, which can be recorded using: o Passive sensors: record reflected solar EMR (e.g., optical, thermal). o Active sensors: emit and record EMR reflected back (e.g., LIDAR, SAR). Optical remote sensing deals with reflected sunlight and is affected by the atmosphere. Spectral signatures help identify materials like vegetation, soil, and water. Radar (SAR) uses complex-valued measurements. Atmospheric remote sensing studies ozone and other phenomena. Examples of key sensors include: EnMAP (hyperspectral), WorldView-2 and Sentinel-2 (multispectral), Sentinel-1 (SAR), and GOME-2/Sentinel series (atmospheric). Artificial Neural Networks (ANNs) History: o Early models were limited (e.g., Perceptron). o Minsky and Papert demonstrated the limitation of Perceptrons for non-linearly separable data. o Backpropagation enabled multi-layer networks and the rise of Deep Learning (DL). Single-Layer Networks: o Can only solve linearly separable problems. o Can not solve non-linear problems such as the XOR problem. Multi-Layer Networks: o Can approximate any continuous function given sufficient hidden layers (Universal approximation theorem). One layer= 1 line on the plot Theoretical capabilities of NN o Boolean function: represented by a network with 2 hidden layers o Continuous function: approximated by a network with 1 hidden layer Deep Neural Networks (DNNs): o DL= NN architecture and loss function + training data + training algorithm+initialization o Convolutional Neural Networks (CNNs)- image analysis (classification, object detection, segmentation), automatic feature extraction, natural language processing o Recurrent Neural Networks (RNNs) -sequential data processing, image captioning and speech recognition. o Generative Adversarial Networks (GANs)- image generation tasks such as super-resolution, image translation and text-to-image generation. AI for Earth Observation (AI4EO) – Key Points AI4EO is not just about classification but also about retrieving physical and biochemical variables. demands high accuracy, traceability, and reproducibility. uses expert domain knowledge and well-controlled data. Data Characteristics multi-dimensional (x-y-z-t-λ), complex-valued, and multi-modal (SAR, LiDAR, hyperspectral.. Training data is often limited. Applications of AI4EO: o Object detection, segmentation and classification (buildings, ships, etc.) o Land use/land cover classification, change detection. o Fusion of multi-modal data (SAR/Optical, 2D/3D). o Image synthesis (e.g. SAR to optical or vice versa). o Cloud removal, atmospheric sensing, climate studies, and social media data fusion. Open Methodological Challenges: o Integrating physics, Bayes and expert knowledge. o Ensuring reasoning, transferability, and uncertainty. o Addressing explainability, ethics and quantum machine learning. o Focusing on Green AI. EORS- Remote Sensing & Earth Observation Remote Sensing & Earth Observation (EO) Fundamentals Remote Sensing: o Definition: Acquiring info about Earth without contact. o Sensing & recording energy, processing, analysis, application. o Technology for obtaining info about Earth & environment. Objectives: Understand Earth via remote data; insights, actions, interventions. Sensors: o Passive: Measures reflected/emitted natural energy (e.g., sunlight) multispectral (few bands), hyperspectral (lot of bands) o Active: Emits own energy; measures reflection (e.g., radar-long waves(micro), LiDAR-short VNIR). Perspectives: In-situ, hand-held, air-borne, space-borne. EO Data: #bigdata, few labels, multi-modal, global, contactless/non-destructive. Electromagnetic Radiation: Crucial for understanding remote sensing. EO Technologies & Missions Optical Sensors: Measure reflected sunlight, various bands-passive o NASA: Landsat. o ESA: Copernicus Sentinel-2, Sentinel-5(p). o DigitalGlobe: GeoEye/WorldView. o Planet: RapidEye/SkySat/Dove. o DLR: DESIS- earth sensing imaging spectrometer, EnMAP-enviro mapping and analysis program. Laser & LiDAR: Active, measures distance, elevation. o ICESat-2. o ESA: Aeolus. RaDAR: Active, radio waves, all-weather. o ESA: Copernicus Sentinel-1. o DLR: TerraSAR-X & TanDEM-X. Combined Missions: Multiple sensors (e.g., ESA's Sentinel-3 & EarthCARE). Mega-Constellations: Swarms of CubeSats, increased temporal resolution, high spatial resolution Resolution Trade-offs-challenges in achieving diff types of resolution: o Spatial: Smallest discernible feature (e.g., Sentinel-2: 10 m px, Maxar: 0.3 m px). o Temporal: Data acquisition frequency (time it visits a site) of the same area (e.g., Meteosat: 15 min). o Spectral: Number/width of bands (e.g., EnMap: 228 bands). EO Applications Environmental Monitoring: Climate change. Agriculture: Vegetation, anomalies, crop yield, plant health. Disaster Monitoring. Climate Modeling: Biogeochemical cycles (H2O, C, N), memory, water balance. Water Resources: Tracking reservoirs, predicting groundwater. Other: Temp (land/sea), sea level, biomass, magnetic field, trace gases. Summary Remote sensing: lot of data, few labels, multi-modal, global, contactless, using ML What can be observed? reflected or emitted electromagnetic radiation What can't be observed? Limitations: cloud cover, spatial resolution constraints, observing subsurface features or achieving high temporal and spectral resolution simultaneously. What do we actually observe? the interaction of electromagnetic radiation with the Earth Which problems do we face? massive data volumes, limited labels, atmospheric effects, data fusion, resolution trade-offs, and complex data interpretation. How to repurpose missions? Existing mission data can be applied to different problems by understanding the sensor characteristics and developing new analysis techniques. Industrial Metrology Engineering Geodesy: Key Concepts Reality Capture: Geometric & semantic data acquisition/modeling of objects/areas. Setting-out: Transfer of geometric plan (from model) to construction site. Monitoring: Measuring object's geometric state over time. Focus: Quality assessment, sensor systems, reference frames. Applications: Facility & construction management, aerospace. Industrial Metrology: Quality Inspection Purpose: Measure shape/dimensions, ensure design compliance. Range: Small to medium, Accuracy: High (µm to mm). Process: Data acquisition, pre-processing, processing target values. Sensor Systems for Industrial Metrology Laser Tracker: 1. Detects reflector-target that reflects the beam back-position detector 2. Measures angle & distance. Time of Flight (TOF): Measures time of signal travel. Range (r): r = c * t / 2 (c=speed of light, t=time). Accuracy: mm. Start impulse->reflection stop impulse, beats counter while gate open= double the range Interferometry: Uses 2 frequencies, measures the difference. Most accurate length measurement. Moving reflector-target Hz & V Encoders (horizontal, vertical): Measures angles using coded glass circle, counting of stripes pattern comparison Accuracy up to 0.1 mm in 40m. 1. Measure arc length b, 2. know radius R 3. calculate angle α 3. Follows moving reflectors. Hz and V motor Measuring Arm: o Carbon fiber, 3 axes, 7 joints,Angular encoders at joints (no) contact o Measuring Volume: 0.5-5.0 m, Accuracy: 10-100 µm. o 3D coordinates: calculated using matrix per joint: T (transform matrix) 3D coordinate = T1 * T2 * T3 … o d: length of the arm segment, a: eccentricity, θ0: offset of the angular encoder, α: orientation of the next rotation axis, θi: angular reading o Triangulation 1D or 2D: ▪ Determine a dist from a ref plane, laser emits-> reflector-> lens->obj and back ▪ Measures distance from light spot position on sensor. Impact of meteorology o Refraction index n: o +1°C => range decreases by 1 µm/m. o +3 hPa => range increases by 1 µm/m.  Bottle neck TOF Triangulation Interferometry Resolution ~0.2mm 5mm/m…1mm/m 0.01microm =10nm Accuracy ~2mm+2mm/km 0.55mm/m…5mm/m 1mium/m…100nm/m=correct hem ▪ Signal: Microwave range (L-band, ~1.5 GHz), modulated code. ▪ Modulated Code: Pseudo-random binary sequence (PRN), spread spectrum. ▪ Satellite s generates code sequence ▪ Receiver r generates same publicly know sequence-replica ▪ Correlate the received satellite signal with the generated replica->signal travel time ▪ 𝑃r^s = c(𝑇r – 𝑇s) Tr-reciever Ts-satelite Satellite Constellations: o GPS: 6 orbital planes, ~4 satellites per orbit, 20,000 km height o Galileo: 3 orbital planes, 8 satellites per plane, 23,000 km height Interoperability and Compatibility: o Interoperability: Signals from different systems used together. L1 E1, L5 E5b o Compatibility: Signals do not interfere, receivers can’t confuse them, o Integrity: provide timely warnings to users when calc position deviates significantly from real o Modulation: distributin of energy over frequency range, diff signal on same range differ in modulation Control Segment: o Reference Frames: Each system has its own (WGS-84, GTRF, PZ-90.11, BTRF), referred to ITRS. o System Time: Realized by atomic clocks in control center. Positioning Precision: o Pseudorange: Meter level. o Phase Measurements/observations: Centimeter level. Orbit Modeling: o Perturbations: Earth oblateness, Moon/Sun tides, radiation pressure. o Challenges: Modeling solar radiation pressure. History o Sputnik I: First artificial satellite, used for positioning with Doppler shift o TRANSIT: First Doppler-based satellite navigation system o NAVSTAR-GPS: Development started in 1973 o NTS-1: First satellite with rubidium clock o NTS-2: First GPS demonstration satellite o  o First GPS Launch: 1978 o Full Operational Capability (GPS): July 1995 o Full Operational Capability (GLONASS): Jan 1996 o Galileo Initial Service Declaration: 15. December 2016 Sustainability: o UN Resolution: Requests sustained development of geodetic infrastructure 2. Key Concepts Pseudorange Signals: Using a code sequence from the satellite and a replica from the receiver to determine signal travel time. Signal Modulation: Using pseudo random noise modulation for robust ranging signals. Visibility of Satellites: Depends on location and time, with more than 100 satellites visible. Reference Frames: Each navigation system realizes its own reference frame, which should align with the ITRS. Interoperability: Different GNSS signals can be used together. Compatibility: GNSS signals are designed to avoid interference. Integrity: GNSS systems provide warnings when position deviates from the real position by more than the specified limit. 3. Applications Plate Tectonics: Measuring continental drift using GNSS. Vertical Plate Motion: Measuring uplift caused by melting glaciers. Earth Rotation: Monitoring variations in the Earth's rotation axis and length of day. Atmospheric Monitoring: Determining water vapor content for weather prediction. Space Weather: Monitoring electron content in the ionosphere. Remote Sensing: Using signal reflections to determine snow depth, vegetation, and soil moisture. Space Service Volume Extension: Using GNSS signals for orbit determination of high-orbiting satellites even Moon. 4. Space Geodetic Techniques Techniques: GNSS, SLR (Satellite Laser Ranging), VLBI (Very Long Baseline Interferometry), DORIS (Doppler Orbitography by Radiopositioning Integrated by Satellite). Goal: Realize an ultra-accurate and long-term stable global terrestrial reference frame. Significance: Metrological basis for monitoring Earth changes, like sea level rise. Fundamental Observatories: Anchor points for combining space geodetic techniques. Genesis Mission: Co-location satellite mission to support consistent technique combination. Gravity Geodesy Overview Physical and Space Geodesy: Focuses on the Earth’s gravity field. Modern Geodesy Pillars: VLBI, ice altimetry, GPS, gravity field, remote sensing, SLR, DORIS. Earth System Processes: Geometry, kinematics, gravity field, rotation. Mass Transport: Monitored via temporal gravity. o Larger impact: pressure, seal lvl, current, ground water, glaciers , ice caps, terrestrial water storage Gravity Field Static Gravity: Patterns of global ocean circulation. Time-Variable Gravity: Mass and gravity changes. Temporal Gravity: Sustained observation of mass transport from space. Mass Change: Observed and monitored by gravity field observations. Essential Climate Variables (ECVs): Monitored through mass change. ECV examples: Surface temperature, sea level, ice mass, soil moisture, terrestrial water storage. Gravity Field Signals: Mountains, deep sea trenches, subduction zones, mid-ocean ridges. Mass and Gravity: Flattening, mountains, irregular mass distribution, groundwater changes. Various signals: Tides, skyscrapers, Earth as sphere Gravity Observation Techniques Terrestrial Data: Heterogeneous data distribution&accuracy, contains high-frequency signals. Altimetric Gravity: Indirect, derived from mean sea surface, covers oceans high freq signals Gravity Satellites: Orbits, orbit differences, acceleration differences SLR, CHAMP, GRACE-FO,GOCE  Satellite-to-Satellite Tracking (SST): High-low and low-low modes. o SST High-low:tracking low orbit sat (orbit affected by gravity)using gps sat,not direct functional of gravity, non o SST Low-low: GPS-> derermine orbits 2 low sat, inter-satelite ranging-measure dist between low low-gravity effec Satellite Gravity Gradiometry: Direct gravity potential, linear observation equation. o Gradiometer-device on sat measuring differences in g accel at different ponts Key observables: Gravity gradients and GPS orbits. Satellite Gravity Missions o CHAMP: (2000-2010). o GRACE: (2002-2017). ▪ Parameters: Orbit inclination 89°, initial altitude 485 km, K-band microwave link. ▪ Payload: K-Band Microwave Ranging System (KBR), accelerometer, GPS receiver. ▪ Measures: Monthly water storage. o GRACE-FO: improvements: Laser interferometer. o GOCE: (2009-2013). ▪ Parameters: Orbit inclination 96.5°, orbit altitude 254.9 km to 225 km. ▪ innovations: Gradiometer, Drag-Free & Attitude Control (DFAC), low altitude. ▪ Payload: Gradiometer, star tracker, GPS receiver, ion propulsion. ▪ mission profile: Orbit lowering, drag compensation. GRACE and Hydrology o Water Balance Equation: P - R - ET = ΔS (Precipitation, Run-off, Evapotranspiration, Storage change). o Measures the storage change directly and is sensitive to groundwater. o Global Mass Redistribution o Long-Term Trends: Water increase/decrease o Freshwater reservoirs: Monitoring of loss of non-renewable freshwater. o Climate-induced trends: Droughts GRACE and Cryosphere o Ice Mass Variation, Sea Level Rise, Sea level changes, Global Sea Level Rise GOCE and Oceanography-Satellite gravity gradiometry. o Mean Dynamic Ocean Topography (MDT): Deviation of real ocean from ideal H = h-N (altimetry - geoid). o Ocean Current Velocities: Derived from GOCE and satellite altimetry. Future Gravity Missions o Science and User Needs: Higher spatial and temporal resolution. o User Priorities: High temporal resolution, short latencies. o New Technologies: Improved inter-satellite ranging, accelerometry, optical clocks. o Satellite Formations: Improved spatial and temporal resolution. o Mission Proposals: e-.motion2, MOBILE, NASA Mass Change Designated Observable (MCDO), ESA Next- Generation Gravity Mission (NGGM), MAGIC. o MAGIC: Mass-change And Geoscience International Constellation. ▪ Joint ESA/NASA Mission: Requirement Document (MRD). ▪ Constellations: Single-pair, pendulum, bender double-pair. o Simulation Results: Bender pairs outperform single-pair concepts. o Temporal aliasing: Dominant error source. o NGGM & MAGIC Status: Pair 1 (GRACE-C), Pair 2 (NGGM). o MAGIC Improvements: Accuracy, spatial/temporal resolution, short latency. Quantum gravimetry: Future Earth observation from space. CubeSats: Networks for geodetic Earth observation. Introduction to earth system modelling Earth System Modeling & Components Earth System Spheres: Atmosphere, Hydrosphere, Cryosphere, Biosphere, Lithosphere. Atmosphere: Gaseous envelope around the Earth. Dynamics described by Navier-Stokes equations. o Navier-Stokes Equations: ρ Du/Dt = ρ (∂u/∂t + u ⋅ ∇u) + ∇p = η∇²u + ℱ. Hydrosphere: Liquid surface and subterranean water (oceans, lakes, rivers). also described by Navier-Stokes Cryosphere: All regions with frozen water (sea ice, ice sheets, glaciers, permafrost). o Dynamics described by Glenn's law for ice sheet shear strain flow (Σ) and stress(σ): Σ = k(T)σ^n. n-mat const Biosphere: All terrestrial and marine ecosystems and living organisms. no fundamental physical eq of motion Lithosphere: Upper layer of the solid Earth, crustal rock, mantle. not included in Climate/Earth System Models. Climate  Climate Definition: Statistics of weather; temporal and spatial averages of temperature, precipitation. o result of dynamics and nonlinear interactions between the five Earth spheres. Nonlinearity: Navier-Stokes equations imply interactions across scales and chaos. o o Chaos: Small initial differences can lead to large final state differences, making prediction difficult. Climate - Attractor: set of all possible states of a system. Climate-predicting the attractor's position due to forcing. Weather: single trajectory on the attractor. Weather- Predicting evolution of a single trajectory from initial condition Climate/Earth System Models Models: Idealized representations of complex reality, used to expand our understanding of climate. Models involve ignoring, distorting, and approximating, but improve understanding. Key Climate Quantity: Temperature Earth's Surface Temperature determined by balance between incoming and outgoing radiation. o C dT/dt = Ein − Eout. C-heat capacity, Equilibrium Climate: When temperature does not change, dT/dt = 0 and Ein = Eout. Milankovich Cycles: Variations in incoming solar radiation on orbital time scales, due to: o Earth's orbit shape (eccentricity). o Angle between Earth's rotational axis and orbital plane (obliquity). o Direction of Earth's rotational axis (precession). Greenhouse Effect Greenhouse gases (H2O, CO2, CH4) absorb long-wave radiation, reducing Eout but leaving Ein mostly unaffected. Increased CO2 from human activity has increased atmospheric CO2. Surface temperatures have increased accordingly. Global Warming Historical warming, effects on extremes, future warming, spatial pattern of temperature and precipitation, effects on sea level detailed in IPCC AR6. Abrupt Changes: Potential state changes of subsystems like polar ice sheets or the Amazon. o Subsystems as potential ‘tipping elements’ with critical thresholds and ‘tipping points’-uncertain o Atlantic Meridional Overturning Circulation (AMOC) likely weaker, but abrupt collapse unlikely before 2100. o Highly nonlinear dynamics of 'tipping elements' are hard to represent in climate models. o Paleoclimate reconstructions and observations are key to understanding. Paleoclimate Evidence: Rapid temperature jumps (up to 16.5K within decades) during the last glacial period; Dansgaard-Oeschger (DO) events as archetypes of abrupt climate change. o Examples of tipping elements include Greenland Ice Sheet, AMOC, Amazon, South American Monsoon Geodetic Earth observation -global terrestrial reference system Geodesy & Earth Observation Geodesy: Science of measuring Earth's state and changes in space and time. o Develops measurement systems and data evaluation methods. o Provides data for Earth system and climate research. o Creates high-precision reference frames. Monitoring: Detection and recording of changes in space and time. Kilometer to mm, decades to second o Requires stable reference. o Reference stability crucial to avoid artifacts. Sea Level Monitoring Satellite Altimetry: Measures distance between satellite and water surface. o (Almost) global coverage since 1992. o Multiple satellites on different orbits. o Measures sea level changes in the order of millimeters per year. Sea Level Determination: Requires accurate distance and satellite orbit measurements. o Regional trends relevant to coastal residents. o Non-uniform thermal expansion and winds contribute to regional variations. Reference System: Variations-> known to at least 1 order of magnitude better than sea level changes (0.1 mm/year). Global Coordinate System Need: Unified, consistent, ultra-precise, long-term stable global coordinate system. o Backbone for referencing and interpreting measurements. Global geodetic observing system-comprehensive basis for earth reliable system monitoring  o Indep observing systems, Globally distributed station network, appropriate observation gom,Earth orientation in space, locally connected GGOS sites Earth-fixed Coordinate System: Each point described unambiguously by 3D coordinates, not changing in time International Terrestrial Reference System (ITRS): Conventional Earth-fixed coordinate system. o Does NOT change over time. o Geocentric System: Origin at Earth’s center of mass. o Axes: x- and y-axis in the equatorial plane, x-axis through Greenwich, z-axis toward conventional terrestrial pole (CTP). o Scale: Meter (SI). o Fundamental for positioning, navigation, surveying, and Earth system research. o Materialized: By globally distributed geodetic observing stations. International Terrestrial Reference Frame (ITRF) ITRF: Realization of ITRS; set of coordinates and changes of observing stations. o 3D coordinate changes determined with ~0.1 mm per year accuracy. Observing Techniques: o VLBI (Radio telescopes): Connection to fixed stars, ITRF orientation in space, Earth rotation, ITRF scale. o GNSS (GPS, GLONASS, Galileo): Cost-effective, high precision, dense network. o SLR (Satellite Laser Ranging): Satellite orbit determination, Earth’s center of mass (ITRF origin & scale). o DORIS: Microwave signals to satellites, homogeneous global distribution. Combination: Needed to determine all parameters (origin, orientation, scale). o Global Distribution: Requires multiple observing stations. ITRF Requirements: Highest precision, long-term stability, and actuality. Time-Dependent Coordinates: Dynamic processes cause continuous changes. Earthquakes can make ITRS unusable. New ITRF Versions: Computed every 5-6 years. Target Accuracy: 1 mm3D position, 0.1 mm/year change of position. Organizations and Institutions International Association of Geodesy (IAG): Coordinates computation of reference frames. International Earth Rotation and Reference Systems Service (IERS): Coordinates activities. o ITRS Combination Centers: DGFI-TUM, IGN (France), and JPL/NASA (USA). DGFI-TUM: Responsible for determining the ITRF, provides ITRF solution DTRF2020. DTRF2020 Latest ITRF solution, combining VLBI, SLR, DORIS, and GNSS data from 1979-2020. Applications & Importance UN Resolution (2015): Emphasizes importance of precise reference frame. Plate Tectonics: Station motions show signature of plate tectonics. Post-glacial Rebound: Vertical station motions show signature. Tide Gauges: Measurements combined with satellite altimetry. Earth Observation: Relies on multiple measurement systems. Global Coordinate System: Fundamental for many applications. o Positioning, navigation, cadastre, GIS, and Earth system research. TUM: Plays key role in realizing global coordinate system, operates Wettzell observatory.