2. Objectives
Describe potentials in using, refining and linking existing and new
Big Data sources in forestry applications, and evaluate the identified
sources of data through a SWOT analysis.
Based on these data, create decision-support models for
efficient, sustainable and value-creating forestry and forest operations.
Demonstrate the potential in keeping and refining information
throughout subsequent operations in the forestry process.
Validation and pilot tests of project results in relevant
environment (hosted by partner companies).
3. Outlook on Big Databases - forest planning
Detailed
digital
elevation
model
Soil
maps
VHR
imagery
Mobile
laser
scanning
Weather
data and
models
Lidar
based
forest
estimates
Road
databases
Forest
machine
data
4. Demonstrated decision support models
Trafficability maps
Operational logging trafficability mapping
Logging trail visualizations and port-harvest quality
control
Models for operational planning of forest operations
Yield and wood property forecasting
Models to improve efficiency in silviculture operations
8. Layout of main extraction routes
Slope and terrain
Volume
Digital terrain model
9.
10. 10 26.6.2019
Logging trail visualizations and port-harvest quality control
2D LiDAR for continous rut depth measurement
11. -15
-10
-5
0
5
10
15
20
25
30
-15 -10 -5 0 5 10 15 20 25 30
Manual
Laser
LiDAR vs manual: per 5m test blocks and
pass
Left LiDAR vs
manual
Right LiDAR vs
manual
Vihti study Kuru study
12. 12 26.6.2019Ala-Ilomäki et. al Harvester CAN-bus
Harvester CAN-bus data for site trafficability
mapping
GPS/Glonass -
positioning
Harvester followed by a
loaded fwd up to five passes
Harvester courtesy of Ponsse Plc
First harvester with and without tracks on two parallel test
tracks
13. 13 26.6.2019Ala-Ilomäki et. al Harvester CAN-bus
Motion resistance, rut depth of harvester
and forwarder on 1st and 2nd pass
17. Models to improve efficiency
in silviculture operations
Plant order tool
Based on harvester data
Actual growth conditions
Root rot frequency
Site index
18.
19. Conclusions
Operational use of trafficability maps and tools to extract main extraction
routes
Digital elevation model spatial resolution critical
Important to act pro-actively to secure data collection
Interesting potential to use Can-Bus data to for trafficability mapping – other
options also availible (drones, machine-mounted equipment, ..)
Big data applications opens a wide range of possibilites for research and
applications – standardisation crucial!
Previous research now possible to implement with Big Data Streams
Forest machine data key for many applications – start collect standardised
data!
Editor's Notes
More cost-efficient operational units (bigger size and better shaped)
Determination of trafficability more accurate and easier
More accurate bordering of the site
Easier to detect places with trafficability difficulties
Learning by using, experience increases utilization
Connection to changing weather
Benefits of integrating map with other map layers
Landings, main strip roads etc.
Decreasing environmental impacts
Smoothing the seasonal variation in roundwood cuttings
Mounted in the back of the forest machine in 45 degree angle
The LiDAR-derived point cloud data were processed with specially designed software developed by Argone Ltd
Position and speed were calculated by recognizing the trees
Wheel ruts were located and tracked by using a simple version of Monte Carlo localization method
The reference terrain level defined based on lowest points and adjusted based on the points located between the ruts
The results are promising even in trials with limited data
Use of steel tracks no problem
BigData applications offer great possibilities: trafficability of vast areas could be mapped annualy, if all harvesters were equipped with the system
Machine learning could be applied to generalize the measurements for similar conditions
The technology is readily available and improvements are more than likely in the future
Tests of larger scale in operational harvesting work planned for autumn 2018