LIDAR TECHNOLOGY AND ITS APPLICATION ON FORESTRY

LIDAR TECHNOLOGY &

IT’S APPLICATION ON

FORESTRY

K.ABHIRAM1318651

Contents IntroductionTypes of Platforms ApplicationsLiterature Review Case studySummaryReferences

2

What is LiDAR ? (Light Detection And Ranging) Lidar  is a remote sensing technology that measures distance by

illuminating a target with a laser and analysing the reflected light.

Why LiDAR ? It is used generating precise and directly geo-referenced

spatial information. Airborne LiDAR systems use 1064nm Lasers produce a coherent light source. It is Active sensor , do not require sunlight, they can be used

either during the day or at night. 3

Figure 1

Lidar is popularly used as a technology used to make high resolution maps, with applications in

Archaeology Geography  Geology & Geomorphology Seismology Forestry Atmospheric physics  Flood Mapping , and Contour mapping Military Mining Transport

4Figure 2

PLATFORMS of LiDAR

1.Earth-orbiting satellites

2.Fixed-wing aircraft, manned and unmanned

3.Rotary-wing aircraft (helicopters)

4.Static ground-based systems (tripods)

5.Dynamic ground-based systems (vehicles)

6.Bathymetric mapping system

1 .

2.

3.

4. 5.

6.

5Figure 3

Parameters of a LiDAR sensor

Repetition rate: LiDAR will pulse at 200,000 times per second.

Scan frequency: how fast the scanner is oscillating

Nominal point spacing (NPS) Scan angle Flying attitude Flight line spacing Across track resolution Along track resolution Swath Overlap 6

Figure 4

LiDAR System Components

1.Laser Scanner2. GPS3. Inertial measurement unit (IMU)4. Computer Processing Resources

7Figure 4

GPS

LiDAR systems use robust dual-frequency receivers and differential post-processing, utilizing a fixed ground reference station.

Lock on the GPS satellites must be maintained and the LiDAR system must stay within 50 miles of the reference base station.

Typically with not more than 3 to 4 cm of error.

A laser scanner has three sub-components: the opto-mechanical scanner, the ranging unit, and the control processing unit.

Laser scanners are distinguished by high timing (range) accuracy, high sampling density, a high degree of automation.

Laser Scanner

8Figure 5

Inertial Measurement Unit (IMU):

Lidar systems contain three inertial gyroscopes. The angular rotations of the sensor from vertical can be measured.

Inertial measurement systems also contain accelerometers to measure the velocity

Computer Processing Resources:

Reliable computer systems are required to ensure that each individual component is performing correctly.

These computer systems must ultimately integrate the component data streams into usable, accurate elevations on the ground.

Data format: LAS format

.

9

Figure 6

Figure 7

Forestry applications require a precise inventory of individual trees

and groups, or "stands" of trees in order to address

Forest management and planning,

Study forest ecology and habitats, Quantify forest fire fuel, and

Estimate carbon absorption.

Application on Forestry

10

Direct measurements include:•Stand Density•Tree Height•Crown Width•Crown Length

Measuring Forests with LiDAR

Estimates include:•Volume•Biomass•Basal Area•DBH (diameter at breast height)

11

Figure 8

Figure 9

Characteristics of LiDAR Data

Discrete Return LiDAR

Visualization of multiple LiDAR returns in a forest canopy, showing first returns from the top of canopy, second returns from forest understory, and third returns near or on the ground

Full Waveform LiDAR

The load increases at about 30 to 60 times.

The opportunities presented by full waveform technology are mostly in the analysis of vegetation density, mapping live versus dead vegetation, forest fuels analysis, and wildlife habitat mapping. 12

Figure 10

Figure 12

Figure 11

Literature Review

Kevin et al.,(2005) studied the horizontal and vertical information of forest from

the LiDAR data.

It is concluded that direct retrieval of canopy height provides opportunities to

model above-ground biomass and canopy volume.

The information also offers new opportunities for enhanced forest monitoring,

management and planning.

Monika et al., (2006) conducted a research with aerial and terrestrial LiDAR to

provide detailed forest inventory characteristics such as canopy heights and

volumes as well as diameter at breast height.

The estimation of Leaf Area Index (LAI) and forest fuel metrics are also

addressed.

  13

Michael et al.,(2008) studied that the number of forest inventory attributes that may be

directly measured with LiDAR is limited.

They present the status of LiDAR remote sensing of forests, including issues related

to instrumentation, data collection, data processing, costs, and attribute estimation.

Nicholas et al.,(2008) assess the forest structure with airborne LiDAR and the effects of

platform altitude . Including three different platform altitudes (1000, 2000, and 3000 m), two scan

angles at 1000 m (10° and 15° half max. angle off nadir), and three footprint sizes

(0.2, 0.4, and 0.6 m).

The comparison was undertaken in eucalypt forests at three sites, varying in

vegetation structure and topography.

 Birgit Peterson et al.,(2010) demonstrates how LIDAR data were used to predict

canopy bulk density (CBD) and canopy base height (CBH) . The LIDAR data were

used to generate maps of canopy fuels for input into a fire behaviour model.14

CASE STUDY-1:Carbon Accounting of Urban Forest in Chennai City using LiDAR Data

Study Areas:1. Guindy Reserve Forest (GRF), 2. Indian Institute of Technology Madras

(IITM), 3. Central Leather Research Institute (CLRI) 4. Anna University (AU) Total area of the study site is 6.67km2

Objectives:Estimation of carbon stock in the center of the city of chennai.To provide more accurate forest biomass estimation

15

Methodology

Lidar data Ortho images

Regression modelling nDSM=DSM-DTM

Object based supervised

classification

Plot level Biomass Canopy Height Canopy cover

Extraction

Tree inventory data

Percentage MAX & MIN heights

Regression Equations

TGB Estimation

Carbon Stock Estimation 16

Field Survey: A Total of 11 square sample plots with the size of

20m were selected on the basis of stratified random sampling.

Each tree coordinates were recorded by using handheld (GPS) unit.

Tree height was measured by using a vertex Hypsometer.

For each tree, stem diameter was measured at 1.3 m above ground with a diameter tape and the species name was recorded.

For Borassus flabellifer height only measured. In each sampling plot of trees, for multi-stemmed trees, bole circumference was measured separately, and summed.

The digital ortho image which was preloaded in the GPS instrument was used for verify the location of the trees.

17Figure 13

Biomass Estimation in Trees The diameter at breast height (DBH) and height of trees were measured, then

both ends of main trunks and length were measured and volume is calculated.

AGB ( above ground biomass) for trees Y = 1.9724x – 1.0717

For the unavailability of the existing equation for Borassus flabellifer trees. Y = 4.5 + 7.7H Where Y = biomass, kg

H= stem height and x =DBH

18

Canopy Cover ExtractionBefore the LIDAR analysis the canopy cover area should be extracted from the urban features.(1) Generating DTM, DEM, and nDSM from ALTM data, (2) Generating Height thresholding image for masking, (3) Segmentation of ortho image, (4) Supervised Classification for extraction of canopy cover

LiDAR Predicted Biomasscanopy densities, mean and percentile heights, and second-

order height statistics . For this study height percentile parameters are developed for biomass estimation.

Regression analysis is the common way to develop the AGB estimation models

The biomass value converted to carbon stock using a conversion factor with the equation Biomass values were multiplied by 0.45 to get carbon storage values of trees .

C= TGB.CF

19

s.no Biomass (Mg)

Trees count Location

1 102.6494 41 IIT2 130.2500 56 IIT3 126.0892 33 GRF4 109.2947 34 GRF5 17.5677 16 CLRI6 47.9295 35 IIT7 63.9100 45 IIT8 95.8963 28 AU9 63.8459 43 AU10 1018.9200 22 CLRI11 85.6047 27 AU

Shows the estimated biomass on the field with site location

20

RESULTS & DISCUSSION

Table 1:

The relationship between field-measured height and LIDAR-measured maximum height (total 438trees)

The field-measured tree height and LIDAR measured height of 438 trees had an R of 0.957 and RMSE of 0.59 m2

21Figure 14

S.NO

NAME AGB (Mg ha-1)

BGB(Mg ha-1)

TGB(Mg ha-1)

CARBON STOCK(Mg ha-1)

1 Guindy RF 42150.2 10959.05 53109.25 23899.16

2 Anna University

8220.029 2137.2 10357.23 4660.75

3 CLRI 4872.509 1226.85 6099.359 2744.70

4 IITM 31847.4 8280.22 40127.62 18057.42

TOTAL 87090.13 22603.32 109693.45 49362.03

Carbon Stock Estimation of Chennai Urban Forest at the all 4 segments

A very good relationship within the LIDAR 75th percentiles and the field measured AGB.

Simple LIDAR metrics such as height percentiles which was derived from canopy heights within plots, gives an impressive capacity to estimate biomass over an urban environment..

22

Table 2:

STUDY AREA: Ahtanum State Forest in Washington State

Fusion of LiDAR and imagery for estimating forest canopy fuels

CASE STUDY-2:

23

Estimation of canopy fuels by using LiDAR data. Canopy fuels are defined as all burnable materials, which include live and dead

foliage, and redundant stem and branch wood located in the upper forest

canopy.

Canopy fuels are important inputs for fire behaviour models that predict crown

fire behaviour and spread . Parameters considered as follows

1. Canopy Height (CH),

2. Canopy bulk density (CBD),

3. Canopy base height (CBH),

4. Available Canopy Fuel(ACF)

Objective:

METHOLODY

Field Data Raw LiDAR Raw Imagery

Data Processing(Fuel

Calculation)Data Processing

(FUSION)Data Processing (ENVI,ArcGIS)

Independent VariablesLiDARmetrics

Imagerymetrics

Regression Analysis

LiDAR Models

LiDAR/Imagery fusion models

Imagery models

CHCBHCBDACF

Dependent Variables

24

25

Results

The regression models organized by canopy fuel metrics by considering the four parameter

of Canopy height ,Canopy base height ,Canopy bulk density ,Available canopy fuel, and

calaculated the RMSE for every regression model of Lidar data,Imagery data, and Lidar +

Imgaery data and compare the values for these data and observed better results in

combination of Lidar and imagery data.

Figure 15

26Figure 16

27

SUMMARY

It gives directly geo-referenced spatial information.

Highly accurate, high-resolution LiDAR data have particular

utility in forest mapping.

By using the LiDAR technology ,the characteristics of forest

can be acquired in a short period .

The first case study shows how to estimate the biomass by using

LiDAR technology.

The second case study shows the accuracy of estimation of

canopy fuels by using LiDAR data.

References Birgit Peterson, Ralph Dubayah, Peter Hyde, Michelle Hofton, J. Bryan Blair, and JoAnn Fites-

Kaufman,(2010) “Use of LIDAR for Forest Inventory and Forest Management Application” Canadian

Journal of Remote Sensing, 29, 650– 657.

Kevin Lima, Paul Treitza, Michael Wulderb, Benoît St-Ongec and Martin Flood,(2005) “LiDAR remote

sensing of forest structure” Progress in Physical Geography 27,1 (2003) pp. 88–106

Michael A. Wulder Christopher W. Bater, Nicholas C. Coops, Thomas Hilker and Joanne C. White

(2008) “The role of LiDAR in sustainable forest management” Remote Sensing of Environment 90:

415–423.

Monika Moskal, Todd Erdody, Akira Kato, Jeffery Richardson, Guang Zheng and David Briggs, (2006)

“Lidar Applications in Precision Forestry”, Journal of Remote Sensing.

28

Muneeswaran Mariappan ,Subbaraj Lingava, Ramalingam Murugaiyan,Vani Krishnan, Srinivasa

Raju Kolanuvada ,Rama Subbu Lakshmi Thirumeni,(2012) “Carbon Accounting of Urban Forest in

Chennai City using Lidar Data”, European Journal of Scientific Research ISSN 1450-216X Vol.81

No.3 (2012), pp.314-328

Nicholas R. Goodwin , Nicholas C. Coops , Darius S. Culvenor, (2008) “Assessment of forest

structure with airborne LiDAR and the effects of platform altitude”, .

Todd L. Erdody, L. Monika Moskal (2010) “Fusion of LiDAR and imagery for estimating forest

canopy fuels”, Remote Sensing of Environment 114 (2010) 725–737

29

THANK YOU

30