Spatial tools for LiDAR based watershed management and forestry analysis

Spatial tools for LiDAR based watershed

management and forestry analysis

Antonello Andrea, Franceschi Silvia, Tonon Giustino and Comiti Francesco

FOSS4G-EU Como 16 July 2015

WHO AM I?

● co-founder of HydroloGIS with Andrea Antonello

● environmental engineer specialized in hydrology, hydraulics and geomorphology

● PhD student of Science and Technology at the Free University of Bolzano (Italy)

● developed scientific models contained in the JGrassTools library in the field of:

– hydrology

– hydraulics

– forestry● OSGeo Charter Member

JGRASSTOOLS

● geospatial library containing modules for:

– vector and raster processing

– geomorphology

– forestry

– mobile mapping connection● it is the core behind the Spatial Toolbox of uDig GIS

● it can be used stand alone using the application Stage (http://bit.ly/stage_downloads): Spatial Toolbox And Geoscripting Ennvironment tool for environmental modelling

JGRASSTOOLS

JGRASSTOOLS

JGRASSTOOLS

JGRASSTOOLS

JGRASSTOOLS

JGRASSTOOLS

WHAT IS LESTO?

● Open Source

● GIS aware

● library dedicated to

● sciences that make use of

● LiDAR data

Developed and maintained by HydroloGIS and the team of prof. Tonon at the Faculty of Science and Technology of the Free University of Bolzano (Italy).

Contains tools for handling high resolution LiDAR data (LAS) and for LiDAR analysis related to forestry.

Utilitie

s

las re

ading

/writ

ing

las sp

atial

inde

x

las in

fo, o

verv

iew

las p

oint d

ensit

y

Rasterlas interpolation

- DTM- DSM

- CHM

Filterslas height distribution

las histogram

las merge

las threshold

Tools

Buildin

g ex

tracti

on

Fores

try a

nalys

is

Inte

nsity

ana

lysis

MODULES OVERVIEW

MODULES OVERVIEW IN STAGE

MODULES OVERVIEW IN STAGE

PREPROCESSING

Packages including the available pre-processing operations are:

● Utilities

– LasIndexer: creates indexes for LAS files

LasIndexer

PREPROCESSING

Packaging including the available pre-processing operations are:

● Utilities


– LasInfo: prints out information of a LAS file/folder

LasInfo

PREPROCESSING


● Utilities



– LasOverviewCreator: creates a shp with overview

LasOverviewCreator

PREPROCESSING


● Utilities



– LasOverviewCreator: creates a shp with overview

– LasPointDensityExtractor: creates a shp with point cloud density on a given grid

LasPointDensityExtractor

PREPROCESSING


● Utilities

– LasIndexer

– LasInfo

– LasOverviewCreator

– LasPointDensityExtractor● Filter

– LasHeightDistribution: analyze the height distribution and categorize the forest type

LasHeightDistribution

LasHeightDistribution

PREPROCESSING


● Utilities

– LasIndexer

– LasInfo



– LasHeightDistribution

– LasHistogram: creates an histogram of the elevation or intensity of all the points in the LAS file

LasHistogram

elevation

LasHistogram

PREPROCESSING


● Utilities

– LasIndexer

– LasInfo




– LasHistogram

– LasMerger: merges all the LAS file contained in a folder in a single one

PREPROCESSING


● Utilities

– LasIndexer

– LasInfo




– LasHistogram

– LasMerger

– LasThresholder: extracts the points with values interval

RASTER

Raster package contains all the available modules to interpolate and create a raster (DTM, DSM) from raw point cloud. The different interpolation algorithms are

● AdaptiveTinFilter: the implementation of the adaptive TIN method of Axelsons

● Las2BivariateRasterMosaic: uses the bivariate function to interpolate a raster from the point cloud and creates a mosaic of TIF

● Las2RasterInterpolator: interpolates a raster from LAS points using the Inverse Distance Weight method

● LasOnRasterMapper: creates a raster by mapping max/min elevation point in each pixel

● LasTriangulation2Dsm: creates a DSM from the triangulation of the point cloud

RASTER

ADAPTIVE TIN

RASTER

ADAPTIVE TIN

RASTER

MAX HEIGHT

BUILDINGS

This package contains a module to extract the vector data of the buildings from a LAS file.

● LasOnDtmBuildingExtractor: based on the identification of the holes in the ground generated by cutting all the points with an elevation on the ground over a given threshold

● the output shapefile can be cleaned from noise data and smoothed on the boundaries

BUILDINGS

BUILDINGS

FLIGHTLINES

Modules to separate different flightlines inside a LAS file:

● FlightLinesExtractor: creates different las files for each of the different flightlines inside the single las

● FlightLinesIntensityNormalizer: normalize intensity values between different flightlines considering the position of the aircraft (x,y,z)

FLIGHTLINES

VEGETATION

Extrapolation of the whole forest biometric data (e.g. forest biomass) can be obtained through two approaches:

● area-based approaches (AB): forest attributes are estimated by relating plot data to ALS data by statistically procedure

● individual tree crown (ITC) approaches: ITC approaches can use both raster CHM and point ALS data and are aimed to detect position and main characteristics of each single tree. Single-tree records can then be aggregated at plot, forest, watershed or regional scale.

VEGETATION

Extrapolation of the whole forest biometric data (e.g. forest biomass) can be obtained through two approaches:

● area-based approaches (AB): forest attributes are estimated by relating plot data to ALS data by statistically procedure

● individual tree crown (ITC) approaches: ITC approaches can use both raster CHM and point ALS data and are aimed to detect position and main characteristics of each single tree. Single-tree records can then be aggregated at plot, forest, watershed or regional scale.

VEGETATION MODULES

The available modules for single tree extraction are based on the identification of local maxima:

● RasterMaximaFinder: identifies local maxima on raster input data

RasterMaximaFinder

RasterMaximaFinder

RasterMaximaFinder

RasterMaximaFinder

RasterMaximaFinder

VEGETATION MODULES



● PointCluodMaximaFinder: identifies local maxima on point cloud input data

PointCloudMaximaFinder



VEGETATION MODULES



● PointCluodMaximaFinder: identifies local maxima on point cloud input data

● WatershedAlgorithm: delineates the crowns of the trees based on raster data

WatershedAlgorithm

WatershedAlgorithm

DATA VISUALIZATION TOOLS

● profiles of the LAS data considering 4 main directions

● position of the measured trees

● position of the extracted trees with highlighted the type of the tree: true positive, false positive




APPLICATION: STUDY AREA

high local variety in forest structure

AURINA VALLEY

VEGETATION:●Norway spruce (Picea abies)●Larch (Larix decidua)●Stone pine (Pinus cembra)

AREA = 10 km2

points classification

points difference DTM

DTM + DSM

points density = 10 p/m2

APPLICATION: LiDAR SURVEY (2012)

12 CIRCULAR PLOTS

RADIUS = 15 m

BUFFER 150 m ROADS

FOREST STRUCTURES:● young● biplane● adult● multilayer

APPLICATION: FIELD SURVEY (2013)

Plot 777

Plot 1223

Plot 485

Plot 683

APPLICATION: RESULTS

method T1: local maxima, raster based

method T11: local maxima, point cloud based



WATERSHEAD MANAGEMENT

WATERSHEAD MANAGEMENT

● GIS-based tool for predicting the magnitude of LW transport during flood events at any given section within a river basin

● two main processes related to wood debris:

– LW recruitment from hillslopes

– LW transport/propagation along the network

SHALSTABmodel

Unstable slopes connectedto fluvial network

CONNECTIVITYmodel

AREAS THAT PROVIDES LW

Channelwidening

Slopeinstability

EXTREME EVENTS

INPUTS AND PROCESSES OUTPUTSLEGEND

LW TRANSPORT DURING FLOODS

SHALSTABmodel


CONNECTIVITYmodel


TREE HEIGHT ANDFOREST STAND VOLUME

CHM and semi-empirical model

Channelwidening

Slopeinstability

EXTREME EVENTS




SHALSTABmodel


CONNECTIVITYmodel


TREE HEIGHT ANDFOREST STAND VOLUME

CHM and semi-empirical model

CRITICAL SECTIONS AND LW VOLUME

Channelwidening

Slopeinstability

EXTREME EVENTS


Volume and dimensionsof available LW

LW propagationField data

monitoring LW transport


JGRASSTOOLS: INPUT

● digital models of the terrain and vegetation: DTM, DSM, FSV

● DTM derived geomorphology attributes: TCA, slope, connectivity, watershed delineation

● extension of the bankfull area: area covered by water during standard flow conditions

● position and dimensions of bridges and dams: field survey or available cadaster

● superficial geology: rock and deposits (erodible)

LW FROM HILLSLOPES

LW PROPAGATION

● identifies the critical section for the transit of LW in the given stream network

● based on the comparison between the length of the logs and channel width

LW PROPAGATION

criticalnon critical

● finalize the implementation of a Particle Swarming Optimizer for automatic calibration of the models

● improve the propagation algorithm to consider also the height of the water in the rivers

● connect the results of the elaboration of LiDAR data for the evaluation of the volume, height and diameter of the logs in each section

FUTURE PLANS

USEFUL LINKShttp://www.jgrasstools.org

http://bit.ly/stage_downloads

Franceschi [email protected]

THANKS FOR THE ATTENTION!

http://www.jgrasstools.org/

Spatial tools for LiDAR based watershed management and forestry analysis

Technology

las files lasinfo

preprocessing packaging

jgrasstools library

lidar analysis

jgrasstools geospatial

preprocessing packages

spatial tools

use of lidar data