Laser Scanning PDF

# GAYZER SCANNING ## BCIT Various Sources of Digital Data - SONAR - Mobile LiDAR - TLS/HDS - UAV Photo (SfM) - (UAV) LiDAR ## Laser Scanning Survey Process - Planning - Data Acquisition - Registration - Modeling - Output (LOD) - (DERIVED Product) ## BCIT What is LiDAR versus "PhoDAR" (SfM) -...

# GAYZER SCANNING ## BCIT Various Sources of Digital Data - SONAR - Mobile LiDAR - TLS/HDS - UAV Photo (SfM) - (UAV) LiDAR ## Laser Scanning Survey Process - Planning - Data Acquisition - Registration - Modeling - Output (LOD) - (DERIVED Product) ## BCIT What is LiDAR versus "PhoDAR" (SfM) - **LiDAR**: Light Detection and Ranging - uses a RANGING Device (Laser – Class1 EyeSafe) - DIRECT OBSERVATION of 3D Coordinate (discrete range & angle solution) - SINGLE Light Ray originating from Scanners Own Coordinate System (SOCS) - Requires DIRECT GEOREFERENCING – Position and Orientation Solution - **"PhoDAR"**: (or photogrammetrically derived Point-Cloud) derived from Structure from Motion - INDIRECT OBSERVATION as 3D Point Cloud is mathematically re-constructed from multiple intersecting Light Rays from multiple vantage points (overlap/multiple photos). - DIM Re-Construction - dense image matching ## BCIT Components of a “LIDAR System” - **LIDAR Unit** - Scans the ground - **Global Positioning System** - Tracks planes x,y,z position - **Inertial Measurement Unit (IMU)** - Tracks Plane Position - **Computer** - Records Data **Laser + GPS + IMU + Computer** **Timing Synchronization is KEY!** ### LIDAR System - **Laser** - Discrete Range & Angle System (SOCS) - Position & Orientation (IMU origin) - **GNSS** - Base Station (Static over known GCP) - Rover (Dynamic/Kinematic) - **IMU** - Captures Vehicle Orientation (R/P/Y) - Knowledge of System LeverArms/Sensor Offsets - Sensor Calibration **Timing Synchronization is KEY!** ## LIDAR System - Emits high frequency infrared laser beams - Laser is made to sweep (using rotating mirror) - Records time to return of each laser pulse - Distance is Calculated w/ Known Speed of Light - Measurements are post-processed along ## TrueView LIDAR Returns - Lidar supports multiple returns (echoes) from a single transmitted laser pulse. - Energy (light) is partially reflected from different surfaces that intercept the beam until it hits a solid target (100% interception of the footprint). - We can use this to restrict our ground filter to "last" or "only/single" returns. ## LiDAR Parameters ### LiDAR for DUMMIES - Repetition Rate - Laser pulse (example: 200 KHz) - Scan Frequency - Scanner Oscillation Rate - Scan Angle - Same as FOV - Flying Altitude - The higher the altitude, the lower the accuracy - Flight Line Spacing - Nominal Point Spacing (NPS) - Point Sample Spacing, the higher the better, but it makes the computations pricey! - Cross Track Resolution - Along Track Resolution - Swath - The coverage area (function of FOV and altitude) - Overlap - needed for higher accuracy ## Beam Divergence - Df = (Divergence * d) + Di ## Usability, continued - Fully populated data attributes - Consistent calibration - Consistency between data layers - No unnecessary artifacts in surface models ## Quality has 3 aspects - Usability - Completeness - Accuracy ## Usability - Report of Survey is complete and correct - Formal metadata are complete and correct - Consistent, correct, and correctly labeled spatial reference framework - Consistent file names and file formats - Usable tiling scheme ## Accuracy - To make bare-earth DEM: - Measure XYZ of points - Classify points as ground or not-ground - Interpolate ground points to continuous surface - me = measurement error - ce = classification error - ie = interpolation error - **DEM error ≈ < (me²+ce²+ie²)** ## Completeness - Complete coverage - Filling gaps requires remobilization, thus is very expensive - Adequate data density - Adequate swath overlap ## Test against GCPs - Look at large-scale shaded-relief images - Too-high points (positive blunders) ## Our QA protocol: 4 analyses 1. Test against photogrammetry 2. Test against ground control points (GCPs) 3. Look at large-scale shaded-relief images 4. CONSISTENCY analysis of swath to swath reproducibility, with completeness inventory - **Extensive automated processing effectively tests for consistent file formats and file naming ** ## Test against Photogrammetry ## LiDAR Calibration ### In-Situ Calibration - Determine corrections to pitch, roll, heading, scan angle, and (possibly) other calibration parameters - Performed by flying over calibration site that has been accurately surveyed using GPS - Calibration site may contain a large, flat-roofed building and runway, taxiway or parking lot surveyed with GPS - Should be performed every time system is re-installed and at regular intervals ## Different scan mechanisms => different scan patterns on ground - Oscillating scan mirror - Zig-zag (or "sawtooth") scan pattern - Rotating polygon - Line scan pattern (parallel lines) ## LIDAR Calibrations - Why? - Why? - Airborne Laser Scanners are a sophisticated systems of systems - GPS - IMU - Scanning mirror (mechanical component) - Each component has inherent errors ## So this does not happen ## Calibration of Lidar Systems - Purpose: - Correct for any systematic errors - Enable achieved accuracy to meet theoretically-expected accuracy - Verify that subcomponents are working within specs - Note: After poor GPS, poor calibration is the #1 culprit in lidar data sets that fail to meet expected accuracies! - **LIDAR Calibrations** - Need to solve for: - Scanner Offset - Pitch - Roll - Heading - Elevation-GPS/TRJ - Order is important as some errors compound! ## Field Calibration Philosophy - Based upon the principle of repeatability: That the seafloor should appear the same - whatever the azimuth - whatever the speed - whatever the motion history - We test this by examining a "Patch" of the seafloor from many different angles and speeds. If the seafloor does not appear the same we try to analyse why. - We decide which unknowns we wish to solve for and then try to come up with paired line geometries in which the mismatch of a target can be explained by only one of the unknowns. ## In-Situ Calibration - Determine corrections to pitch, roll, heading, scan angle, and (possibly) other calibration parameters - Performed by flying over calibration site that has been accurately surveyed using GPS - Calibration site may contain a large, flat-roofed building and runway, taxiway or parking lot surveyed with GPS - Should be performed every time system is re-installed and at regular intervals ## What does Pitch look like? ## Pitch Trigonometry ## Visualizing a Pitch Error - Determined from bi-direction and fully overlapping flightlines - CrossSection Profile cut in ALONG TRACK direction (NADIR) ## The Roll ’Magic Triangle‘ ## Visualizing a Roll Error - Determined from bi-direction and fully overlapping flightlines - CrossSection Profile cut in ACROSS TRACK direction (SINGLE SWATH) ## What does Heading look like? ## Visualizing a Heading Error - Determine at outer edge of overlapping flightlines flown @ same direction - CrossSection Profile cut in ALONG TRACK direction (thru OUTER overlap) ## The Heading ‘Magic Triangle’ ## Features of Laser Scanning - Captures 3D position data of any point - Dense data acquisition - Non-contact → No need for instrumentation of sensors on structure - Ability to capture data for structures that are not easy to access - Reasonably accurate for many applications - Versatile - Efficient - Easy to use ## How Laser Scanning Works - **Phase Shift** - Measures change in phase of multiple sinusoidal laser pulses to determine time and distance - **Time of Flight** - Measures time it takes a laser pulse to travel to target and back to determine distance ## Operation - Set up equipment - Scanner, computer - Position targets as needed - Scan - Use software for post-processing - Visualization - Stitching/Registration - Meshing ## Applications - **Historical Documentation** - Palace Museum in the "Forbidden City" in Beijing - **Construction Inspection** - Survey - **The Big Dig** - Survey - Closeout drawings - **Forensics** - I-35 W bridge in Minnesota - Riegl LMS Z420 - **As-built documentation** - Retrofit ## RELATIVE POSITIONING - BLK360 – P SERIES | | BLK360 | RTC360 | P-Series | |-------------|---------------|--------------|--------------| | Imaging Speed | < 1min | < 1min | > 6min | | ScanSpeed | 360'000 pts/sec | 2'000'000 pts/sec | 1'000'000 pts sec | | Level of Detail (Max Resolution) | Low (6mm @ 10m) | Medium (3mm @ 10m) | High (0.8mm @ 10m) | | Env. Robustness | IP54/open mirror | IP54 / closed mirror | IP54 /closed mirror | | Temp Range | +5 to +40 | -5 to +40 | -20 to +50 | | Field Registration | SW based | VIS based | Surveying Procedures | | Range (typical/max) | Short (20/60m) | Medium (40/140m) | Long (70/270m) | | 3D data quality | Good | Better | Best | Accuracy aaportunity - to discuss not only RANGE but also ANGULAR and point out how POOER Misleadgin Spec sheets are..... ## Topographic ## Scanning Applications - As-Built Surveys/Topo - Concept Visualization - Asset Management - Construction Assessments ## Reason 1: Capture Everything ## Reason 2: Accuracy & Confidence ## Reason 3: Visualization ## AS-BUILT: A drawing or model that directly reflects the current conditions of a building ## Concrete Wall Assessments - Scan Survey Wall Surface Model 500,000 Points - Conventional Survey Wall Surface Model 400 Points - Point accuracy vs modeled accuracy - Generalization vs detail ## Scanning Advantages - Reason 1: Capture Everything - Reason 2: Accuracy & Confidence - Reason 3: Visualization ## AS-BUILT: A drawing or model that directly reflects the current conditions of a building ## THE TERRESTRIAL SCANNING PROCESS - PROJECT PLANNING - TARGET PLACEMENT AND MEASUREMENT - SCANNING - REGISTRATION - MODELING - CONVERSION TO CAD - RENDERING AND VIDEO ## PROJECT PLANNING - IS THIS THE RIGHT TECHNOLOGY? - SCOPE OF THE WORK - LEVEL OF COMPLETENESS - LEVEL OF DETAIL - LEVEL OF ACCURACY - DEFINE THE DELIVERABLES - CYCLONE (POINT CLOUD OF MODEL) - ACAD / MICROSTATION (2D PLAN OR 3D MODEL) - VIDEO - OTHER POINT CLOUD FORMATS ## GATHER SITE INFO - RISKS - TRAFFIC, CONSTRUCTION, RESTRICTIONS - REGULATIONS, ACCESS AND TIME, REMOTE SITE - POWER SUPPLY - WEATHER - ONLY OPERATES DOWN TO OC - BCIT SECURITY - REQUIRES SPECIAL FORM FOR UNSUPERVISED STUDENTS BECAUSE OF CLASS 2 LASER ## Level of Detail - Scope of Work - What needs to be extracted - Architectural - Structural - MEP - Fixtures - Furniture - LOD Specification - How closely the model will represent the real-world ## IN FIELD PRACTICES - NAME AND MEASURE TARGET POSITIONS WITH TOTAL STATION - SCAN 2 GO - SCAN FROM - SCAN ALL - NO DROPS - 3/3/3 - 60/40 @TRANSITIONS - FIND THE BLOG ## REGISTRATION - DURING SCAN CREATE HIGH DENSITY SCAN OF EACH TARGET - CAN BE PRE-CODED AS TARGET ## NAME AND MEASURE TARGET POSITIONS WITH TOTAL STATION - CONVERT MEASUREMENT TO PROPER COORDINATE SYSTEM RT TO CYLONE ## SCAN SETUPS/KEYPLAN ## Summary - Scanning provides significant cost savings over the lifespan of a project - Point clouds are data rich thus eliminating the need for repeated site visits. - Rapid date collection results in less time in the field. - Accurate datasets equate to less rework. - Clash detection analysis decreases costly onsite design revisions. - Scanning provides ac accurate record of existing site conditions. - Record as-built drawings are often found to be unreliable. - Survey Calculations - Engineering Designs - Construction Management - Scanning is the foundation for any reliable BIM. - An accurate model is a valuable tool to be used for: - Planning - Scheduling - Estimating - Analyzing - Managing ## CONCLUSIONS - NAME TWO SCANNER MODELS - LIST 4 STEPS DESCRIBED TODAY IN THE SCANNING PROCESS - LIST ISSUES WHEN GATHERING SITE INFO? - WOULD IT BE POSSIBLE TO REGISTER POINT CLOUDS WITHOUT CONTROLS? - WHAT IS THE SCAN ANGLE WITH THE CYRAX SCANNER? - HOW MUCH OVERLAP IS REQUIRED ## CYCLONE FIELD 360. WORKFLOW OVERVIEW. - Cyclone Register 360 – Step 1: Import - Cyclone Register 360 – Step 2: Review and Optimize - Cyclone Register 360 – Step 3: Finalize - Cyclone Register 360 – Step 4: Report ## Isolating Points ### Limit Box - Why do we need to isolate points? - Limit box easy to use, adjustable - Limit box does not effect performance - Limit box parallel to view - Set a pick point - Set Limit Box by Cursor - Click and drag to size - Adjust edges - Limit box may be toggled on and off ## Isolating Points ### Fence - Fence allows multiple shapes - Fence is only 2D - In order to clip in 3D requires two steps - Fence affects performance - Possible to select inside or outside fence - Possible to Hide selected - Possible to assign selected to a layer. ## Isolating Points ### Layers - Like all CAD Cyclone supports layers - Layer Properties - New to create new layer - Set Current Bold Name sets active - Assign- Assign selected elements to that layer - Visibility, Select ability, Assign colour - Works with Points and Modeled objects ## Isolating Points ### Region Grow - Great way to select points without having to create a fence - Low impact on computer performance - Directly create Patch, Cylinder, Sphere, Surface, Pipe Run ## Affect of modeling on accuracy - When we make a patch in cyclone it take thousands of points and averages them into a plane. - Name some of the issues that you have run into in modeling. - Unless you model you don't know how to scan. - The Incas where building to higher accuracy than we can currently measure. ## Accuracy in HDS - All functions that convert a point cloud to an object are an estimate - Patch, elbow, box, Point cloud subject to noise - Angle of incidence - Colour of object - Distance - Edges - Multiple Scans ## Accuracy in HDS - **Aerial LiDAR** - Up to 1 - 40 ppm - 10cm XY, 5cm Z - **Aerial LiDAR Helicopter** - 40 to 400 ppm - 5cm XY, 3cm Z - **Aerial Photography** - Potentially 1100ppm - Relates to distance - 3cm XYZ - **First generation terrestrial** - 1cm X,Y,Z - Patch creation 4mm - **Scan Station 2 C10** - Position 6mm, Distance 4mm - Patch creation 2mm - **Terrestrial Photogrammetry** - Relates to distance - 0.3mm X,Y,Z - 2 pixels X the photoscale ## How to Handle Handles - Multi-Select the Objects - Use shift single select - Zoom in on the handles - Press hold shift and drag one handle to the other and it will snap, - This may rotate the object - Just move one handle - Press hold shift + CTRL and drag and it moves the whole object ## NUT SHELL - SLAM is a technique used to build up a map within an unknown environment or a known environment while at the same time keeping track of the current location. ## SLAM – Multiple parts - Landmark extraction - date association - State estimation - state update - landmark update - There are many ways to solve each of the smaller parts ## What is SLAM? - The problem has 2 stages - Mapping - Localization - The paradox: - In order to build a map, we must know our position. - To determine our position, we need a map! - SLAM is like the chicken-egg problem - Solution is to alternate between the two steps. ## Hardware - Mobile Robot - Range Measurement Device - Laser scanner – CANNOT be used underwater - Sonar – NOT accurate - Vision – Cannot be used in a room with NO light ## The goal of the process - The SLAM process consists of number of steps. - Use environment to update the position of the robot. Since the odometry of the robot is often erroneous we cannot rely directly on the odometry. - We can use laser scans of the environment to correct the position of the robot. - This is accomplished by extracting features from the environment and re observing when the robot moves around. ## Extended Kalman Filter - An EKF (Extended Kalman Filter) is the heart of the SLAM process. - It is responsible for updating where the robot thinks it is based on the Landmarks (features). - The EKF keeps track of an estimate of the uncertainty in the robots position and also the uncertainty in these landmarks it has seen in the environment. ## Laser & Odometry data - Laser data is the reading obtained from the scan. - The goal of the odometry data is to provide an approximate position of the robot. - The difficult part about the odometry data and the laser data is to get the timing right. ## Landmark Extraction - Once we have decided on what landmarks a robot should utilize we need to be able to somehow reliably extract them from the robots sensory inputs. - The 2 basic Landmark Extraction Algorithms used are Spikes and RANSAC ## Landmarks - Landmarks are features which can easily be re-observed and distinguished from the environment. - These are used by the robot to find out where it is (to localize itself). ## The key points about suitable Landmarks - Landmarks should be easily re-observable. - Individual landmarks should be distinguishable from each other. - Landmarks should be plentiful in the environment. - Landmarks should be stationary. ## RANSAC (Random Sampling Consensus) - This method can be used to extract lines from a laser scan that can in turn be used as landmarks. - RANSAC finds these line landmarks by randomly taking a sample of the laser readings and then using a least squares approximation to find the best fit line that runs through these readings. ## Data Association - The problem of data association is that of matching observed landmarks from different (laser) scans with each other. - Problems in Data Association - You might not re-observe landmarks every time. - You might observe something as being a landmark but fail to ever see it again. - You might wrongly associate a landmark to a previously seen landmark. ## Overview of the process - Update the current state estimate using the odometry data - Update the estimated state from re-observing landmarks. - Add new landmarks to the current state. ## Algorithm - Nearest Neighbour Approach - When you get a new laser scan use landmark extraction to extract all visible landmarks. - Associate each extracted landmark to the closest landmark we have seen more than N times in the database. - Pass each of these pairs of associations (extracted landmark, landmark in database) through a validation gate. - If the pair passes the validation gate it must be the same landmark we have re-observed so increment the number of times we have seen it in the database. - If the pair fails the validation gate add this landmark as a new landmark in the database and set the number of times we have seen it to 1. ## Final Review – Open Areas - There is the problem of closing the loop. This problem concerned with the robot returning to a place it seen before. The robot should recognize this and use the new found information to update the position. - Furthermore the robot should update the landmarks found before the robot returned to a known place, propagating the correction back along the path.