Project SLOPE Forest Monitoring

Project SLOPE
1
WP 2 – Forest information collection and
analysis
San Michele all’Adige
19-21 Jan 2015

SLOPE WP 2 – Task 2.1
Alberto Lupidi, PhD
Remote sensing and multispectral analysis
Remote Sensing Department
Flyby S.r.l.
19-21 Jan 2015

Task 2.1: participants
• CNR
• Coastway
• Flyby S.r.l. (Task Leader)
• TreeMetrics
19-21 Jan 2015

Task 2.1: general description
1. Define a methodology to obtain a description of the scenarios using
available remote sensing data (From satellite, UAV and on ground
instrumentation)
2. Define how to realize a more complete forest inventory
AIMs:
Flyby
Define the approach
to monitor tree
growth and health in
mountainous
environment (Using
different vegetation
indexes)
CoastWay/Treemetrics
Define the approach to
monitor the forest using
UAV and on ground
sensors
CNR/Flyby
Define the approach to fuse
heterogeneous information
(derived by satellites or other
instrumentations)
Participants Role
19-21 Jan 2015

Main Results
• Definition of VI from sensed data
• Estimation of biological parameters from VI
• Analysis of parameter with growing level of
detail (Satellite -> UAVs -> Laser scanner)
• Utilizability
19-21 Jan 2015

Issues, delays and actions
• Case study chosen, with a posterior inspection, revealed too
young/thin trees
• The alternative (Gortahile forest) was not present in Sat data
(geographical mismatch)
• Delays from due date (30 Aug 14) to achieve a suitable solution.
• Action Taken: utilization of a combination of Ballyredmond and
Trentino data to show the effectiveness in VI utilization at different
scales and with different sensors
19-21 Jan 2015

Working on a case study
IRELAND
MONTESOVER
19-21 Jan 2015

RapidEye satellite imagery
19-21 Jan 2015

Vegetation Indices
redNIR
redNIR
NDVI
+
−
=
redRE
redRE
NDRE
+
−
=
NDVI
NDRE
CCCI =
Normalized Difference Vegetation Index
Used for vegetation/wood detection. Saturates
in dense wood areas
Normalized Difference Red-Edge Index
Can be utilized to estimate chlorophyll levels
/health status of trees.
Unaffected by saturation
Canopy Chlorophyll Content Index
Proposed as an alternative to measure
variations in biomass, chlorophyll and
nourishment (Nitrogen content)
19-21 Jan 2015

NDVI / NDRE
NDVI
Oct - Jul
NDRE
Oct - Jul
19-21 Jan 2015

CCCI
October 2012 July 2013
19-21 Jan 2015

Chlorophyll estimation
October 2012 July 2013
19-21 Jan 2015

NDVI – Chl , UAV sensed data
Trentino dataset – enhancement of detail and granularity
19-21 Jan 2015

From forest database to stem characterization
Trunks analysis with terrestrial laser sensor
19-21 Jan 2015

From data to information: the workflow
19-21 Jan 2015

Conclusions
• Development of an integrated wood monitoring
system
• Definition of a chain processing for data from
different platoforms
• Relation between VI indices and biomass productivity
index
• Toward very detailed information to tree-level for
woodcutting operators
19-21 Jan 2015

Alberto Lupidi, PhD
Flyby s.r.l.
Corso Ferrucci 77/9, 10138 Torino, Italy
Via Puini 97, 57128 Livorno, Italy
www.flyby.it
Thanks
19-21 Jan 2015

Species recognition by foresters

Post Processing can be time consuming as time is spent looking
for the targets
New RTK unit will include all open areas in the DSM

The same GCP location identified in 6no. Photographs is difficult to tie dow

19-21 Jan 2015
First Step - After IrishTests
Identification ofTheTest Sites in Piscine & Montesover

19-21 Jan 2015
Second Step – Submit Flight Plans to ENAC
(ItalianAviationAuthority)

WP2 Overall Progress ofTasks - Submitted and Pending deliverables
Task 2.1- Remote sensing and multispectral analysis FlyBy declared
complete 14/10/2014
Task 2. 2 - UAV data acquisition and processing Coastway declared complete
14/01/2015
Task 2.3 - On-field digital survey systems Treemetrics declared complete
16/12/2014
Task 2.4 Harvest simulation tool based on 3D forest model Declaration
required
Task 2.5 - Road and Logistic Planning Boku declaration required to be
discussed
WP2 meeting required involving all Contributors
19-21 Jan 2015

WP2 - Submitted and Pending Deliverables
Task 2.1- Remote sensing and multispectral analysis FlyBy declared complete
14/10/2014
Task 2.1- Remote Sensing and Multispectral analysis FlyBy Plan for Austrian
Test
Task 2. 2 - UAV data Acquisition and Processing Coastway Revise document
Plan for Austria
Task 2.3 - On-field Digital Survey Systems Treemetrics Revise Document Plan
for Austria
Task 2.4 Harvest simulation tool based on 3D forest model deliverable required
Task 2.5 - Road and Logistic Planning Boku deliverable required
Questions to be asked in meeting;
Will the same reports be produced for the Austrian site?
19-21 Jan 2015

WP2 Criticalities
Prepare Flight Plan and Site
Criticalities for WP2 Forest Information Collection and Analysis to be a
success are
FlyBy, Coastway , Treemetrics, CNR have shared data this must
continue.
Graphitec to reveal how this data is being used
Boku to share the scientific data with all WP2 contributors
Equipment to be upgraded in line with current developments for
Austrian test
19-21 Jan 2015

WP2 Recommendations
Data to be discussed between the Survey Group and the
Scientists
A larger area of the forest to be scanned to complete a
full model
The current model to be revealed with a viewer for all
parties
19-21 Jan 2015

T2.3-TLS data and analysis
Kick-off Meeting 8-9/jan/2014
Delive- rable
Number
61 Deliverable Title
Lead benefi-
ciary number
Estimated
indicative
person- months Nature 62
Dissemi-
nation level 63
Delivery date 64
D2.01 Remote Sensing data and analysis 7 14.00 R PU 8
D2.02 UAV data and analysis 4 11.00 R PU 8
D2.03 TLS data and analysis 9 13.00 R PU 10
D2.04 Harvest simulation tool based on 3D forest model 1 20.00 R PU 15
D2.05 Road and logistic simulation module 10 12.00 R PU 13
Total 70.00
19-21 Jan 2015

D2.03TLS data and analysis
D2.03) TLS data and analysis: This deliverable will
contain a report on TLS data collected, the
methodologies and algorithm to extract needed
information and the generated output information.
[Completed on time in month 10] (Updated
following edits 16/12/2014)
19-21 Jan 2015

Overview of ForestryAnalysis
Typical Forest inventory workflow and monitoring
19-21 Jan 2015

Types of Samples (plots and transects)
Sample Plots
• Stratum sizes
• Variability inside the strata
• Inventory costs and accessibility that
might vary between strata
• Target accuracy
Inventory Planning is Essential in order to:
1. Ensure efficient and effective data collection
2. Minimise cost
3. Optimise accuracy
19-21 Jan 2015

Treemetrics Field Gathering App
The advantage of in field application:
Bluetooth recording with callipers
2 way communication between office and
field using GPRS or Wifi.
Easy navigation in field for operator
19-21 Jan 2015

Field Measurement
TLS Height recording hypsometer DBH
recording
19-21 Jan 2015

Field Data Analysis
Dendrometric study the measurement of the
various dimensions of trees, such as
their diameter, size, shape, overall volume, etc.
Using the TLS the following data is available:
Stem profile
Diameter at Breast Height (DBH)- DBH v Height,
DBH v Crown Size
Tree Volume
Quality indicators (Straightness)
19-21 Jan 2015

TLS Analysis for Plot 2
Stem 3D profile in Autostem
19-21 Jan 2015

Demonstration
Overview of the Demo area
The first demonstration area is
located in Piscine (Province of
Trento, Italy). This demo area has
4.18 ha of mixed age forestry. It is
composed mainly of Norway
spruce (Picea abies) mixed
intimately with other secondary
species: Scots pine (Pinus
sylvestris), European larch (Larix
decidua), silver fir (Abies alba) and
Pinus cembra.
19-21 Jan 2015

Demonstration
A second demonstration area
located in Montsover (Province
of Trento, Italy), was initially
proposed for study. However
TLS sample has been not
recorded in this site, due that the
trees for extraction had not been
marked in the field when field
data collection was performed.
Therefore, the combination of
TLS and aerial/satellite remote
sensing cannot be performed.
19-21 Jan 2015

Spatial Analysis
Spatial information is essential in order to know how the forest parameters are
spatially distributed in the forest stand. Therefore, the combination of forest
parameters and remote sensing must be used to study how the dendrometric
parameters are distributed.
Typically, the spatial analysis is based on the following steps:
1. Location of the forest area (Area of analysis).
2. Pre-processing of the image
3. Zoning for analysis.
4. Delineation of the forest canopy (Forest productive area).
5. Tree detection (if applicable).
6. Forest stratification based on tree characteristics (if applicable).
7. Inventory Plan (Samples plot location).
8. Generalization based on field data (parameters distribution and final
stratification).
19-21 Jan 2015

Demonstration
EO Photo UAV Photo
EO image UAV image
Source Bing Maps UAV
Resolution 0.5 m 0.1 m
Bands RGB RGB-Nir
19-21 Jan 2015

Demonstration
DTM Slope
DTM generated from LIDAR was available for the demo area. This has a
resolution of 1m. From the DTM the slope has been calculated in order to
have a better understanding of the demo area.
19-21 Jan 2015

Demonstration
Digital Height Model
19-21 Jan 2015

Demonstration
Productive Area
EO DATA UAV DATA with DTM
DHM from UAV data allows to detect scrub areas
19-21 Jan 2015

Demonstration
Stratification
19-21 Jan 2015

Demonstration
Tree Counting
DHM from UAV data allows to detect scrub areas
19-21 Jan 2015

Demonstration
Tree Counting – EO data
Plot
Manual
counting
Trees
Detected
(matches)
False
detections
Over-
segmented
Double
crowns
Matching
Validation
rate
1 20 15 1 9 2 75%
2 24 20 3 4 2 83%
3 21 14 0 2 3 67%
4 24 15 5 2 4 63%
5 20 16 1 0 2 80%
Total 109 80 10 17 13 73%
19-21 Jan 2015

Demonstration
Tree Counting - UAV with DTM
Plot
Manual
counting
Trees
Detected
(matches)
False
detections
Over-
segmented
Double
crowns
Matching
Validation
rate
1 20 20 1 2 0 100%
2 24 21 2 4 0 88%
3 21 20 1 1 0 95%
4 24 21 1 3 0 88%
5 20 20 0 0 0 100%
Total 109 102 5 10 0 94%
19-21 Jan 2015

Demonstration
y = 0.3573x2
- 2.2107x + 5.5136
R2
= 0.0335
0
1
2
3
4
5
6
0 1 2 3 4 5
Crown diameter (m)
Volume
(m3)
y = 137.46x2
- 11.411x + 2.239
R2
= 0.0372
0
1
2
3
4
5
6
0.0 0.1 0.1 0.2
ExRG index
Volume
(m)
No correlation with volume using EO data (Vegetation index and Crown
Diameter)
19-21 Jan 2015

Demonstration
Good correlation with UAV DHM (DBH prediction from Crown Diameter)
(Volume prediction from Height*Basal Area)
y = -0.0069x2
+ 0.4969x - 0.0193
R2
= 0.7064
0
1
2
3
4
5
6
0 5 10 15
Predicted Height * Basal Area (m3)
Volume
(m3)
y = 0.618x - 1.3349
R2
= 0.6215
0
1
2
3
4
5
6
0 2 4 6 8 10
Crown diameter (m)
DBH
(m)
19-21 Jan 2015

Demonstration
19-21 Jan 2015

Technical Meeting 19-20-21/jan/2015
Task 2.4 - 3D Modelling for
harvesting planning
19-21 Jan 2015

PART 1
- Task goals, timeline and partecipants
PART 2
- Alpha Version of the 3d Harvesting Planner: integrating the data coming from
task 2.1,2.2,2.3;
- first interactive planning tools;
- first version of cable crane setup.
PART 3
- Live Demo of the 3D Forest Model
Outlook
19-21 Jan 2015

Aims
Deliverable 2.4: Harvest Simulation Tool based on 3D forest
model / Software platform to visualize harvest processing
Develop methodologies and tools to fully describe terrain and
stand characteristics, in order to evaluate the accessibility for
and efficiency of harvesting technologies in mountain forests.
19-21 Jan 2015

TimeLine
Start Month: 7 – August 2014
End Month: 15 –April 2015
Deliverable: Harvest simulation tool based on 3D forest
model
Total MM: 20
Task leader: GRAPHITECH;
Participants: CNR, COAST, BOKU, GRE, FLY, TRE
19-21 Jan 2015

Participants role
GRAPHITECH (10MM): Task Leader. It has in charge the development of
software tool for representing the virtual 3D environment of the mountain forest
as well as the of the virtual system on mobile and machine-mounted displays.
Finally it will be involved into the development of the solution for interactive
cableway positioning simulation.
CNR (1MM): Definition of the “technology layers” and methodologies to
coordinate tree marking with the subsequent harvesting operations.
COASTWAY (2MM): Provide the input model taken from UAV survey data
combined with the information of task 2.1 and 2.3
19-21 Jan 2015

Participants role
BOKU (2MM): it will be involved into definition of the “technology layers” (i.e.
harvest parameters) for cableway positioning.
GREIFENBERG (1MM): Acting as final user in order to simulate the behaviour of
its own machines into the virtual system.
FLYBY (1MM): Provide the input model for the virtual system combining the
information of task 2.1, 2.2 and 2.3.
TREEMETRICS (2MM): Development of the Forest Warehouse for mountain
forestry and support the deployment of the virtual system. Provide input data
taken form terrestrial laser scanning analysis.
19-21 Jan 2015

Functionalities
• Forestry measurements estimations:
The platform will allow the combination of accurate tree profile
information with up to date remote sensing data.
• Interactive tools for cableway positioning simulation.
• Definition of the “technology layers” (i.e. harvest parameters);
Technological layers show technical limitations of machines and
equipment on different forest areas.
• Deployment of the virtual system on mobile and machine-embedded displays.
19-21 Jan 2015

Two levels of abstraction
1St Level: 2D map
accessing of forest and
logistic information
including:
Cadastal dat, Volume of
timber, accessibility of
area.
Where available, the
system allow access to the
SLOPE forest information
system.2nd Level: 3D map
accessing of forest with
information and
visualization of the single
tree and stems profile.
Cable Crane positioning
and tools to plan
harvesting area
19-21 Jan 2015

Platform Core
Canopy surface model Laser scanner point cloud Aerial & satellite images
19-21 Jan 2015

Integrating Data
Using the remote data (Satellite, UAVs orthophotos and digital surface model) combined
with on field information (TLS), each single tree feature will be segmented including its
deducted geometric properties.
Task 2.1
Task 2.2
Task 2.3
3D forest model
Virtual 3D
environment
19-21 Jan 2015

Input Data
TEST CASE- PILOT A – Monte Sover
Round-up of the data coming from deliverables 2.1.,2.2,2.3
Flyby
Coastway
Treemetrics
19-21 Jan 2015

3D Harvesting Planner
Alpha version of the 3d model of forest
WebGL Based : HTML5 canvas – no need for plugin, run in almost any modern
browser
CesiumJS: extend an Open-Source library for GEO-Visualization on virtual globe
(measuring tools, point clouds, cabke crane, vehicle tracking, etc.)
Serving spatial-data through Tile Map Service and Web Map Service compliant
server
19-21 Jan 2015

Visualization of Digital Surface Model of the test case pilot area
PISCINE AND SOVER:
DSM derived from UAV processed images
19-21 Jan 2015

Visualization of Orthophoto RGB and NIR aerial images of the test case
pilot area
PISCINE AND SOVER:
This data are derived from UAV processed images
19-21 Jan 2015

Single Tree visualization and stems profile
Visualization of the trees segmentation and single tree stem profile. The
data are coming form Treemetrics surveys with terrestrial laser scanner
19-21 Jan 2015

Planning Tools for measurements
User could draw a polygonal area ora a line that will be used to take
measurements useful to setup a landing zone or any other operational
area
19-21 Jan 2015

Cableway positioning simulation and cable crane following Greifenberg
guidelines and suggestion
19-21 Jan 2015

Cableway positioning simulation and cable crane following Greifenberg
guidelines and suggestion
• Typology of Cableway used by Greifenberg in alpine/mountainuous
scenario
• Operational lane of carriage/lifting
• Catenary Simulation
• Constrains on real operational scenario for cableway construction
19-21 Jan 2015

Cable Crane Setup alpha-phase : Catenary Simulation
Setup a new cable crane by set the points of start and arrives of the cable
19-21 Jan 2015

Future Roadmap
- Improve cable crane setup and configuration
- Dynamic load of external layers
- GPS vehicle tracking
- Show properties for each tree
- Terrain analysis tool
- User interface as described within D.1.2 Human Machine Interfaces Definition
- Vegetation Index and time series analisys visualization for deeper
forest/biomass analysis
- Complete Integration with the Forest Warehouse devolped by Treemetrics
- Integration with the Forest Information System developed by MHG
- Technology layers definitions on cable harvesting.
- Integration with Road and Logistic Planning
- Virtual Tree Marking
19-21 Jan 2015

Open Discussion
19-21 Jan 2015

Thank you for your attention
GIULIO PANIZZONI
giulio.panizzoni@graphitech.it
Fondazione Graphitech
Via Alla Cascata 56C
38123 Trento (ITALY)
Phone: +39 0461.283393
Fax: +39 0461.283398
19-21 Jan 2015

Technical Meeting
2015
Project SLOPE
1
T 2.5 – Road and Logistic planning
Trento, 19th-21st January, 2015
19-21/jan/

Index
2
Task objectives
Work done so far
SLOPE logistic model
Data model
Next steps
Contact info

1. Task objectives
3
Task objectives:
Identify and analyze logistics elements within the forest and their
characteristics for site locations and flow allocation decisions
Integration of the data with the global forest model
Build and validate an Optimization model to allocate landings and transport
wood to mills and processing plants
To be developed from M8 to M13
Includes development of “D2.05 Road and logistic simulation module”
Due to Month 13.
Partners involved: all
ITENE (leader), GRAPHITECH, CNR, BOKU, FLY
Technical Meeting
19-21/jan/2015

1. Task objectives
4
LOGISTIC OPTIMIZATION MODEL:
The goal is to determine an optimal forest logistic
network to respond future demands.
The approach would determine:
Allocation of each buffer to the customer considering lowest
supply costs
Which storage yards to open and how much timber will flow
via these yards during damage events, i.e. when too much
timber is produced within a certain period
Technical Meeting
19-21/jan/2015

1. Task objectives
5
The forest supply
network consists of
several nodes:
Roadside stocks
(buffers, B).
Central storage
yards (S).
Customers (C),
representing a
saw mill, paper
mill or biomass
plant.
Technical Meeting
19-21/jan/2015

2.Work done so far
6
Task activities
by partner: ITENE
(Task leader)
- Understanding the forestry supply chain and logistic processes.
- Identifying relevant logistic locations
- Defining characteristics of logistics elements
- Extensive bibliographic research
- Coordinate the task activities
- Integration with the global model
- Elaboration of ‘Data Model’ for the Task integration
- Choosing a real scenario with real data
BOKU
- Participate in the common meetings and issues
- Understanding the forestry supply chain and logistic processes.
- Providing feedback of deliverables and other documents.
- Developing the logistic model
- Programming the logistic model
- Validating model in real scenario
CNR
- Participating in the common issues
- Identifying relevant logistic locations
- Providing feedback of deliverables and other documents
GRAPHITECH
- Participating in the common meetings and issues
- Providing feedback of deliverables and other documents
FLYBY - Participating in the common meetings and issues
Done
In progress
Future task
Technical Meeting 19-21/jan/2015
Technical Meeting
19-21/jan/2015

2. Work done so far
7
Templates to collect some data and bibliographic information
Technical Meeting
19-21/jan/2015

2. Work done so far
8
2 possible scenarios to test the model:
Italy - Alps area (maybe in Trento) CNR
(Still in contact. TBC)
Austria - “Austrian State Forest” BOKU
(Still in contact. TBC)
Technical Meeting
19-21/jan/2015

2.1. SLOPE logistic model
9
DECISION VARIABLES
xijk Volume to be transported from i to j for assortment k
yi 1, if storage yard i should be opened
zik 1, if a buffer i is available for delivery for assortment k
DATA VARIABLES
sik Volume of wood at i for assortment k
cijk Transport costs from i to j for assortment k
chik Harvesting costs at i for assortment k
csmik Costs for manipulating of assortment k at storage yard i
djk Demand at j for assortment k
fi Fixed costs for opening storage yard i
Minimum turnover of storage yard i for assortment k
Maximum capacity of storage yard i for assortment k
Technical Meeting
19-21/jan/2015

10
OBJECTIVE FUNCTION:
To supply timber at the lowest supply costs (harvesting, transportation
and fixed and variable costs at storage yards):
Harvesting costs for
(chi) differ because o
reasons, like amoun
each buffer
f several
t of timber,
tree volume, extraction distance
etc.
The sum of transport costs are calculated as the
ca lat
volume flow multiplied by costs for each assortmentmultipli(k) and each used route.
If storage yards are opened (yi=1),
fixed costs (fi) per month occur.
The sum of variable processing
costs at storage yards are
lcu ed the volume flow
ed by costs for
manipulating for each assortment
(k).
Technical Meeting
19-21/jan/2015

11
CONSTRAINTS:
The demand of assortments k at the customer destinations j has
to be satisfied within a specific range.
On one hand the resource limits at the harvesting plots (buffers)
must be kept, but on the other hand we do not want left overs of
assortments like saw logs or pulp wood.
For storage yards the wood flow must be balanced.
For considering a safety stock, a minimum turnover via storage
yards can be determined.
Depending on the capacity of the storage yard, a maximum
turnover per month can be defined.
The decision variable for the wood flows has to be nonnegative.
The decision variables if a storage yard (yi) will be opened and if
delivery should be executed from the buffer (zik) can only take
the value 0 or 1.
Technical Meeting
19-21/jan/2015

12
SUPPLEMENTARY CALCULATIONS:
Harvesting costs:
Road transport costs:
Cijk Supply costs from origin (i) to destination (j) [EUR/m3]
tijk
nik
sik
Drive time from origin (i) to destination (j) for assortment k [h]
Loading time for assortment k [h]
Unloading time for assortment k [h]
Operational delay time for assortment k [h]
Cost of transport vehicle per hour per road category r [EUR/h]
Road charge from i to j [EUR]
Calculated number of trips needed to transport the total volume
from i for assortment k
Volume of wood at i for assortment k
Costs for construction and maintenance of forest roads per road
category r [EUR/m³]
Cother
Other costs e.g. shifting costs for trucks, pre-concentration of raw
material [EUR/m3]
Chik Harvesting costs at i for assortment k [EUR/m3]
Csaw
Cyarder
Cprocesso
r
tvik
Cost of felling with chain saw per hour [EUR/h]
Cost of extracting with tower yarder per hour [EUR/h] Cost
of processing with processor head per hour [EUR/h]
Average tree volume at i for assortment k [m³]
distik Average extraction distance at i for assortment k [m]
slik
Cmount
Cdmount
Average slope at i for assortment k [%]
Costs for mounting the cable line [EUR/m3]
Costs for dismounting the cable line [EUR/m3]
Technical Meeting
19-21/jan/2015

2.2. Data model
13
BUFFER (B)
STORAGE YARDS (S)
CUSTOMERS (C)
B/C
B/S
S/K
á
ASSORTMENTS (K)
tlk
tuk
tdk
B/S/K
B/C/K
S/C/K
B/K
C/K
S/C
Technical Meeting 19-21/jan/2015
! "
# $%
& %! !
$
' %
Technical Meeting
19-21/jan/2015

3. Next steps
14
Implementation of the basic optimization model and test with theoretical data
Confirmation of real scenario and stakeholders involved
Collect additional requirements to the optimization model coming from real
scenario stakeholders, other inputs (e.g. MHG)
Get real scenario input data for the model
Model tested with real data from real scenario
Data model to be integrated to SLOPE data model
Draft of deliverable to be sent to all partners
Feedback integration and final deliverable
Parallel session
To discuss
deadlines and
options to get
a real scenario
Technical Meeting
19-21/jan/2015

4. Contact info
15
Emilio Gonzalez
egonzalez@itene.com
Loli Herrero
dolores.herrero@itene.com
Technical Meeting
19-21/jan/2015

eting
2015
16
Thanks for your attention!!!
Technical Meeting
19-21/jan/2015

Project SLOPE Forest Monitoring

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (18)

Similar to Project SLOPE Forest Monitoring

Similar to Project SLOPE Forest Monitoring (20)

More from SLOPE Project

More from SLOPE Project (19)

Recently uploaded

Recently uploaded (20)

Project SLOPE Forest Monitoring