Advance surveying lab (CE-605) - Unit 4 Short Notes

UNIT 4: ADVANCED SURVEYING TECHNOLOGIES & APPLICATIONS

Module 1: Advanced Positioning & Navigation Systems

Global Navigation Satellite Systems (GNSS) - Advanced Concepts

• Real-Time Kinematic (RTK): Provides real-time, high-accuracy (cm-level) positioning using a fixed base station that broadcasts correction data to a rover via radio link or network (NTRIP). Requires initial integer ambiguity resolution.

• Post-Processed Kinematic (PPK): Similar to RTK but corrections are applied in post-processing software. More flexible for areas with poor real-time communication, often used with UAVs.

• GNSS Augmentation Systems:

  • SBAS (Satellite-Based Augmentation System): e.g., WAAS (USA), EGNOS (Europe). Uses geostationary satellites to broadcast correction signals, improving accuracy to ~1-2 m.

  • PPP (Precise Point Positioning): Uses precise satellite orbit and clock corrections from global networks. Achieves cm to dm accuracy without a local base station, but has longer convergence time.

• Multi-Constellation GNSS: Utilizes signals from GPS (USA), GLONASS (Russia), Galileo (EU), BeiDou (China). Benefits: Increased satellite availability, improved geometry (lower PDOP), better reliability in urban canyons/obstructed areas.

• Sources of GNSS Errors & Mitigation:

  • Ionospheric Delay: Mitigated using dual-frequency receivers or models (e.g., Klobuchar).

  • Multipath: Reduced by using choke ring antennas, proper site selection (away from reflective surfaces), and advanced signal processing.

  • Ephemeris & Clock Errors: Corrected via broadcast ephemerides or precise products from IGS.

  • Atmospheric (Tropospheric) Delay: Modeled using standard atmospheric models or estimated in processing.

[!TIP] Exam Focus: Distinguish RTK (real-time, radio link) vs. PPK (post-processed, for UAVs). Know that multi-constellation improves PDOP (Position Dilution of Precision).

Total Station & Robotic Surveying - Advanced Operations

• Reflectorless (EDM) Measurement: Uses laser pulse to measure distance to non-cooperative targets. Limitations: Shorter range, affected by surface reflectivity/color, atmospheric conditions, and beam divergence.

• Robotic Total Station: Motorized, can be controlled remotely via data collector. Features target acquisition (auto-tracking) and search function. Requires clear line-of-sight and stable tripod.

• Integration with Data Collector: Seamless field-to-office workflow. Data collector runs survey software, stores points, codes, and attributes directly in the field.

• Advanced Traverse & Resection: Instrument software performs least-squares adjustment for traverse networks and free station (resection) setups, providing coordinate residuals and accuracy estimates.

[!TIP] Common Pitfall: Reflectorless measurements fail on dark/absorbent surfaces or in heavy rain/fog. Always verify with a prism for critical points.

Module 2: Modern Data Acquisition Platforms

Unmanned Aerial Vehicles (UAVs) for Surveying

• Regulations & Flight Planning: Must comply with national aviation authority (e.g., DGCA in India). Flight planning software (e.g., Pix4Dcapture, DroneDeploy) sets waypoints, altitude, and front/side lap.

• UAV Platforms:

  • Multirotor: Vertical take-off/landing (VTOL), ideal for small, complex sites. Limited endurance.

  • Fixed-Wing: Longer range/endurance, efficient for large, open areas. Requires runway or catapult launch.

• Key Flight Parameters:

  • Altitude: Lower altitude = higher Ground Sample Distance (GSD) = higher detail.

  • Overlap/Sidelap: Typically 70-80% forward, 60-80% side lap for reliable Structure from Motion (SfM) processing.

  • GSD (Ground Sample Distance):

$$ \text{GSD} = \frac{\text{Sensor Height} \times \text{Sensor Width}}{\text{Focal Length} \times \text{Image Width}} $$

(in meters/pixel).

• Ground Control Points (GCPs) & Check Points (CPs):

  • GCPs: Pre-surveyed targets (high-accuracy GNSS) used to georeference the entire model.

  • CPs: Independent points used only for accuracy assessment (RMSE calculation). Must be well-distributed and not used in processing.

Airborne and Terrestrial Laser Scanning (LiDAR)

• System Components: Laser (emitter/receiver), Scanner (mirror mechanism), IMU (Inertial Measurement Unit for orientation), GNSS (positioning).

• Data Acquisition Principles:

  • Point Density: Points per square meter (ppm). Determined by pulse rate (kHz), scan angle, and flight speed/altitude.

  • Pulse Rate: Higher kHz = more points, denser data, larger file size.

• Mobile vs. Terrestrial LiDAR:

  • Mobile LiDAR: Mounted on moving vehicle (car, boat). Integrates GNSS/IMU for trajectory. Efficient for corridor mapping (roads, railways).

  • Terrestrial Laser Scanning (TLS): Stationary, tripod-mounted. High accuracy, high density. Used for structures, buildings, complex sites. Requires multiple scan stations and target-based registration.

• Target Placement & Georeferencing: For TLS, retro-reflective targets are placed around the site. Their known coordinates (from GNSS/total station) are used to merge multiple scan stations into a single coordinate system.

[!TIP] Exam Focus: GSD formula is critical. For LiDAR, know that point density is a key specification. GCPs are for georeferencing, CPs are for accuracy validation—never mix them.

Module 3: Data Processing, Modeling & Integration

Photogrammetric Processing (Structure from Motion - SfM)

• Standard Workflow:

  1. Image Import & Alignment: Software detects keypoints, matches images, computes camera positions and sparse point cloud.

  2. Sparse Cloud: Initial 3D point cloud from camera geometry. Used for initial quality check.

  3. Dense Cloud Generation: Uses depth maps to create high-resolution 3D point cloud.

  4. Mesh/Texture Creation: Generates continuous 3D surface (mesh) and applies image colors (texture).

  5. Orthomosaic & DSM: Orthorectified 2D image map (orthomosaic) and Digital Surface Model (DSM - includes objects like buildings/trees).

• Software Tools: Agisoft Metashape, Pix4Dmapper, RealityCapture. Interface varies but core steps are similar.

• Accuracy Assessment: Import GCPs and CPs. Compute RMSE (Root Mean Square Error) in X, Y, Z. Target RMSE depends on project specs (e.g., ±5 cm for engineering).

Point Cloud Processing & Analysis (LiDAR & Photogrammetry)

• Visualization & Management: Use specialized software (e.g., CloudCompare, Autodesk Recap, LAStools). Handle large datasets via octree or level-of-detail (LOD) rendering.

• Point Cloud Classification: Automated/semi-automatic process to label points:

  • Ground: Bare earth.

  • Vegetation: Low, medium, high vegetation.

  • Buildings/Structures.

  • Noise/Outliers: Remove erroneous points (e.g., from birds, atmospheric effects).

• Filtering & Clipping: Remove unwanted areas (e.g., vegetation to expose ground for DEM), slice at specific heights, isolate objects.

• Export Formats:

  • LAS/LAZ: Standard binary format for LiDAR point clouds (LAZ is compressed).

  • RCS/RCP: Autodesk Recap format.

  • ASCII (TXT/CSV): Simple x,y,z format.

Data Fusion and Integration

• Integrating Point Clouds with Raster & Vector Data:

  • Overlay orthomosaic (raster) on point cloud for context.

  • Import survey points (vector) to check alignment or add control.

• Co-registration of Multi-Source Datasets: Aligning datasets from different sources (e.g., TLS point cloud with UAV DSM) using ** ICP (Iterative Closest Point)** algorithm or common targets/features.

• Creating Hybrid 3D Models: Combine high-accuracy TLS data for structures with broader UAV coverage for context.

[!TIP] Common Pitfall: SfM fails on featureless surfaces (snow, water, white walls). Add targets or use LiDAR. Always classify point clouds before volume calculation to separate ground from objects.

Module 4: Applications, Deliverables & Quality Assurance

Volume and Cut/Fill Calculations

• Methods:

  1. TIN-based (Triangulated Irregular Network): Creates a surface from 3D points. Volume between two TIN surfaces (e.g., existing vs. design) computed via prismoidal formula for each triangle.

  2. Cross-Sectional: Volumes calculated from area differences between successive cross-sections (e.g., for roads, canals).

  3. Grid-based (Digital Model): Uses regular grid (DSM/DEM). Volume = sum of (height difference × grid cell area).

• Prismoidal Formula (for a single triangular prism): For two parallel surfaces with areas $$\displaystyle A_1 $$ and $$\displaystyle A_2 $$ and mid-area $$\displaystyle A_m $$, separated by height $h$:

$$ V = \frac{h}{6} (A_1 + 4A_m + A_2) \boxed{} $$

*Applied iteratively over a TIN.*

• Reporting: Report net volume (cut - fill), gross cut, gross fill. Include accuracy estimate (e.g., ±2% of volume).

As-Built Surveys and 3D Modeling

• 3D Model Creation: From point clouds (TLS/UAV), generate 3D meshes of existing infrastructure (buildings, bridges, pipes).

• 2D Plan Deliverables: Extract from 3D model:

  • Contours (from classified ground points/DEM).

  • Profiles & Cross-sections (by slicing 3D model along alignment).

• Clash Detection & Deviation Analysis (BIM): Compare as-built 3D model to design BIM model. Software (e.g., Navisworks) identifies geometric conflicts (clashes) and deviations beyond tolerance.

Quality Assurance / Quality Control (QA/QC) in Advanced Surveying

• Field Check Surveys: Conduct independent, higher-accuracy surveys (e.g., total station) at CPs to validate primary data (UAV/TLS).

• Statistical Analysis of Residuals:

  • RMSE (Root Mean Square Error):

$$ \text{RMSE} = \sqrt{\frac{\sum_{i=1}^{n} (e_i)^2}{n}} $$

where $$\displaystyle e_i $$ is the residual (difference between measured and check value).

*   **Standard Deviation (σ):** Measures spread of errors.

*   **Mean Error (Bias):** Indicates systematic offset.

• Documentation Standards: Maintain metadata (sensor settings, processing parameters), processing logs, and accuracy report (including RMSE for GCPs/CPs).

• Accuracy Standards: Follow ASPRS (American Society for Photogrammetry and Remote Sensing) or NSPS (National Society of Professional Surveyors) standards for different products (e.g., Class 1 LiDAR, orthomosaic NMAS).

[!TIP] Exam Focus: Prismoidal formula is essential. RMSE is the primary accuracy metric. QA/QC is not optional—it's a mandatory documented process.

Module 5: Emerging Trends & Comprehensive Project Workflow

Introduction to Related Technologies

• Ground Penetrating Radar (GPR): Sends electromagnetic pulses into ground. Detects subsurface utilities, voids, layer interfaces. Data processed as time-slices or 3D volumes. Accuracy depends on soil conductivity.

• Mobile Mapping Systems (MMS): Integrated system with multiple sensors (LiDAR, cameras, GNSS/IMU) mounted on a moving vehicle. Captures street-level 3D data efficiently for asset management.

• AI & Machine Learning: Used for automated feature extraction from point clouds (e.g., classifying power lines, trees, buildings) and images (object detection). Reduces manual editing time.

End-to-End Project Workflow

1. Client Requirements & Feasibility: Define accuracy, deliverables, budget, timeline. Assess site access, permitting, and technology suitability.

2. Technology Selection: Based on:

   • Scale: Large area → UAV; small detailed → TLS.

   • Accuracy: Sub-cm → Total Station/GNSS RTK; dm-cm → UAV PPK; m-level → SBAS.

   • Deliverables: 3D model → LiDAR/UAV; contours → GNSS/total station.

   • Budget & Time: UAV is fast/cost-effective for large areas; TLS is slow but high-detail.

3. Integrated Field Planning: Combine methods if needed (e.g., UAV for topography, TLS for building). Design GCP/CP network. Plan flight paths/scan stations.

4. Data Processing Chain: From raw data (images, point clouds) to final deliverables (orthomosaic, DEM, 3D model, volume report). Includes QA/QC checkpoints at each stage.

5. Final Deliverable & Reporting: Package all products (raw data, processed data, reports, metadata) per client/specification requirements.

[!TIP] Exam Strategy: For scenario-based questions, justify technology choice based on accuracy, scale, and deliverable. Always mention GCPs/CPs and QA/QC in any project plan.