Scribd Logo
Upload

Welcome to Scribd!

  • 43 views

    3D Change Detection: Presented By: Shagun Garg (183040064)

    Uploaded by

    Puneet garg
    This document presents a methodology for detecting 3D changes between lidar point clouds from 2009 and 2013. The objective is to classify changes in buildings and vegetation by comparing lidar-derived digital surface models (DSMs) and digital terrain models (DTMs) between the two years. Slope and volume thresholds are used to classify changes into four types: (1) new buildings, (2) building changes, (3) demolished buildings, and (4) no change. Validation with 2013 building polygons achieved 85-95% accuracy, with errors mainly from construction sites and low 2009 point density.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    3D Change Detection: Presented By: Shagun Garg (183040064)

    Uploaded by

    Puneet garg
    43 views13 pages
    This document presents a methodology for detecting 3D changes between lidar point clouds from 2009 and 2013. The objective is to classify changes in buildings and vegetation by comparing lidar-derived digital surface models (DSMs) and digital terrain models (DTMs) between the two years. Slope and volume thresholds are used to classify changes into four types: (1) new buildings, (2) building changes, (3) demolished buildings, and (4) no change. Validation with 2013 building polygons achieved 85-95% accuracy, with errors mainly from construction sites and low 2009 point density.

    Original Description:

    3d change detection

    Original Title

    RS_GIS_mini

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document presents a methodology for detecting 3D changes between lidar point clouds from 2009 and 2013. The objective is to classify changes in buildings and vegetation by comparing lidar-derived digital surface models (DSMs) and digital terrain models (DTMs) between the two years. Slope and volume thresholds are used to classify changes into four types: (1) new buildings, (2) building changes, (3) demolished buildings, and (4) no change. Validation with 2013 building polygons achieved 85-95% accuracy, with errors mainly from construction sites and low 2009 point density.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    43 views13 pages

    3D Change Detection: Presented By: Shagun Garg (183040064)

    Uploaded by

    Puneet garg
    This document presents a methodology for detecting 3D changes between lidar point clouds from 2009 and 2013. The objective is to classify changes in buildings and vegetation by comparing lidar-derived digital surface models (DSMs) and digital terrain models (DTMs) between the two years. Slope and volume thresholds are used to classify changes into four types: (1) new buildings, (2) building changes, (3) demolished buildings, and (4) no change. Validation with 2013 building polygons achieved 85-95% accuracy, with errors mainly from construction sites and low 2009 point density.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    You are on page 1of 13

    3D Change Detection

    Presented by :
    Shagun Garg
    (183040064)
    Introduction

    Fig (A) : Vertical Profile of trees (2009) ; Fig.(B) : Builidings profile (2013)
    Souce : Dong. et.al(2018)

    2
    IDEA....
    Trees generally have higher
    slope than buildings.

    Fig 2. Principle of nDSM


    https://www.stadtentwicklung.berlin.de

    3
    Objective
    • Type I: New buildings built after removing
    medium/high vegetation.

    • Type II: New buildings built on bare earth


    or low vegetation, or atop
    existing buildings.
    • Type III: Demolished or damaged buildings.

    • Type IV: Existing buildings with little or no


    changes.
    Fig.3. Types of Change.
    Souce : Dong. et.al(2018)

    4
    Fig.4. Lidar point cloud in LAS format for (A) 2009 and (B) 2013. Souce : Dong. et.al(2018)

    2009 2013
    Points density : 2 Points /m2 25-30 points/m2
    No. of Returns: 4 5
    Classes: 5; ground, noise, water, 8; Ground, low, medium, high
    overlap, reserved. vegetation, building,
    noise,water,reserved 5
    Problem...
    • Objective was to compare the Building and
    vegetation classes in the year 2009 and 2013.
    • In the year 2013 we have two different classes
    - vegetation and buildings
    • But in the year 2009 , the two classes
    (vegetation and Buildings ) are not classified.
    • Using 2013 classified points we will classify
    the 2009 unclassified points.

    6
    Methodology
    2009 Lidar Points 2013 Lidar Points

    2009 DEM 2009 DSM 2013 DSM 2013 DTM

    Filter out G
    2009 nDSM dDSM 2013 nDSM
    and B Points

    Change in Volume.
    Height of building

    Slope threshold

    7
    Thresholds..
    Slope Threshold

    Maximum Slope : 51.11 degree


    Threshold : 51.11 + 2 degree
    =53.11 degree

    If mean slope of 2009 DHM in


    same polygon is greater than
    Threshold(St) then objects
    derived in 2009 are vegetation.
    OR
    Fig.5. Histogram of zonal mean rooftop slope
    If Si>St : Type 1 change
    from 2013 data

    8
    Thresholds..
    Volume Threshold

    Vt = 2.2 * Building footprint area


    Where , 2.2 is the minimum height of a residential building [Standard]
    Vt =Volume Threshold
    Vij = Volumeric change of ground objects from 2009 to 2013
    = DSM (2013) – DSM(2009)
    Vij = dDSM.
    Where , Vij =Change.

    Change > -Threshold Else


    Change > Threshold
    Vij > - Vt
    Vij > Vt TYPE-4
    TYPE-3
    TYPE-2 9
    Methodology
    2009 Lidar Points 2013 Lidar Points

    2009 DEM 2009 DSM 2013 DSM 2013 DTM

    Filter out G
    2009 nDSM dDSM 2013 nDSM
    and B Points

    Slope Raster 2013 Building


    Slope
    Threshold (St)
    Use 2013 BA as Use 2013 BA as
    zones to calc Si zones to calc Vij Vol. Threshold (Vt)
    =2.2 *(building area)
    Check Si , St, Vij, Vt

    NO NO
    NO Type 4
    Si > St Vij > Vt Vij > -Vt

    YES YES YES

    Type 1 Type 2 Type 3 10


    Validation
    1112 Building points
    from 2013 Lidar data

    Fig 6. Map of Building Types . The 2009 DHM is used as backdrop


    11
    Souce : Dong. et.al(2018)
    Conclusion
    • 85-95 % Overall accuracy
    • No clear threshold between demolished and
    damaged.
    • Can be used only in areas where safety
    standards are followed.
    • Errors are mainly caused due to buildings
    under construction , buildings under tree
    canopy, and relatively low density of 2009
    Lidar points.
    12
    THANK YOU

    13

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy