3rd Technical Meeting - WP2
0 likes647 views
This document summarizes work from Project SLOPE on collecting and analyzing forest information using remote sensing techniques. It describes using a UAV to acquire RGB and multispectral imagery of a test site in Annaberg, Austria. Terrestrial laser scanning was also used to generate digital terrain models, detect individual trees, and analyze tree characteristics like volume estimations. Field measurements were taken and compared to remote sensing data. The integration of remote sensing with field data helped improve forest inventory and management techniques.
3rd Technical Meeting - WP2
- 1. Project SLOPE 1 WP 2 – Forest information collection and analysis
- 2. Review Meeting 19 – 21 January Corcaigh WP2 Slope Project WP2.2 UAV Data Acquisition and Processing Corcaigh January 19th- 21st 2016 Partners: Coastway /CNR / Treemetrics / FLYby
- 3. Review Meeting 19 – 21 January Corcaigh WP2 Request for Permission - AustroControl
- 4. Review Meeting 19 – 21 January Corcaigh WP2.2 October 2015 weather conditions Precipitation The probability that precipitation will be observed at this location varies throughout the month. Precipitation is most likely around October 31, occurring in 58% of days. Precipitation is least likely around October 5, occurring in 53% of days. Probability of Precipitation at Some Point in the Day in October Snow The likelihood of snow falling is highest around October 31, occurring in 12% of days. Probability of Snow Fall Being Reported in a Given Day in October Moderate rain is the most severe precipitation observed during 64% of those days with precipitation. It is most likely around October 2, when it is observed during 37% of all days. Light rain is the most severe precipitation observed during 11% of those days with precipitation. It is most likely around October 15, when it is observed during 6% of all days.
- 5. Review Meeting 19 – 21 January Corcaigh WP2.2 Permission Granted 17- September 2015
- 6. Review Meeting 19 – 21 January Corcaigh WP2.2 Permission Granted 17- September 2015 Licence to fly fitted to all equipment
- 7. Review Meeting 19 – 21 January Corcaigh Annaberg Austria – Test Site
- 8. Review Meeting 19 – 21 January Corcaigh WP2.2 Ground Markers Installation of survey targets and static GPS observations over several days
- 9. Review Meeting 19 – 21 January Corcaigh WP2.2 Calibration of Cameras Calibration of Multispectral Camera
- 10. Review Meeting 19 – 21 January Corcaigh WP2.2 flight Plan 120m above ground level
- 11. Review Meeting 19 – 21 January Corcaigh WP2.2 Image Log from UAV
- 12. Review Meeting 19 – 21 January Corcaigh WP2.2 Canopy Model on Lidar Test Test combine with scan data. Test 3 Trento Flight all data sets recorded Satellite, UAV, TLS, Marking
- 13. Review Meeting 19 – 21 January Corcaigh WP2 Proposed Trials View models using QT Software
- 14. Review Meeting 19 – 21 January Corcaigh “Beyond the State of the Art” The Advantages of Multispectral data over RGB & IR multiSPEC 4C - Ultra precise 4-band accuracy The multiSPEC 4C provides image data across four highly precise bands - Green, Red, Red- edge and NIR - with no spectral overlap. In addition, its upward-facing irradiance sensor automatically compensates for sunlight variations, resulting in unparalleled reflectance measurement accuracy. Technical features Resolution 4 sensors of 1.2 Mp Ground resolution 10 cm/px (@100m) Sensor size 4.8 x 3.6 mm per sensor Pixel pitch 3.75 um Image format RAW (Tiff)
- 15. Review Meeting 19 – 21 January Corcaigh WP2 – Coastway Involvement 23.5months RGB Data enables you to determine the species and density of growth Budget Man Months Total Returned Bal Remaining WP 1 4 4.9 -0.9 WP 2 11 11.67 -0.67 WP 3 0 0 0 WP 4 0 0.75 -0.75 WP 5 2 1.75 0.25 WP 6 0 0 0 WP 7 4 3.4 0.6 WP 8 2 1.36 0.64 WP 9 0.5 0.63 -0.13 23.5 24.46 y WP Budgets f
- 16. Review Meeting 19 – 21 January Corcaigh WP2 Problems experienced & Solutions Bureaucracy can be a big burden on accessing forestry Weather can be a problem such as high winds / snow Access permits are required / permission fly varies from region to region IAA, CAA, ENAC , Austro Control & the Forest Wardens. GPS quality can be poor the accuracy varies from 250mm to several meters but suffecient for forestry new developmemts of an RTK UAV does away with the need for GCP’s Access to the mountain forest was restricted so a small clearing was located to enable launch of the UAV.
- 17. Review Meeting 19 – 21 January Corcaigh WP2 Beyond the state of the Art - Forestry Industry Surveying Forestry with UAV’s opens up many fields which include. Forestry Management for planting, thinning and harvesting Disease identification Forest Fire monitoring Search and rescue Landslide monitoring
- 18. Review Meeting 19 – 21 January Corcaigh WP2 Advantages of using Multispectral Camera Growth and yield of the forestry crop is the result of its interaction with environmental factors, soil conditions and availability of water and nutrients in the soil. With the availability of crop-health detecting sensors, it its now possible to obtain indicator kinetics such as NDVI, biomass, chlorophyll rate, leaf area index, water stress, flowering ,on a much smaller scale than before. Multispectral imagery can determine the health, Species, of trees The multispectral imagery can help determine haul routes by the reflection from different soil types. The imagery can be used to determine growth rates over several seasons and help determine where thinning is required. Coastway have received a lot of interest from Forestry companies in the UK & Ireland regarding forestry management.
- 19. Review Meeting 19 – 21 January Corcaigh WP2.2 Multi spectral data - NDVI - Normalized Difference Vegetation Index - Annaberg
- 20. Review Meeting 19 – 21 January Corcaigh Questions Thank you for your attention endanolan@coastway.net
- 21. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Introduction Work Package Number 2 Start date M4 Work Package Title Forest information collection and analysis Activity Type RTD Participant Name (only contributing partners) TreeMetrics Person-months p-m RP2 July-October 2015 2,90 p-m total planned in DoW 10,00 10,00 person-months has been spent in this workplace since the start of the project The participation of Treemetrics in this work package is now completed.
- 22. Kick-off Meeting 8-9/jan/2014 TLS data and analysis This deliverable contains a report on TLS data collected, the methodologies and algorithm to extract needed information and the generated output information. T2.03TLS data and analysis Introduction
- 23. Kick-off Meeting 8-9/jan/2014 -Montsover (Province of Trento, Italy) - Norway spruce (Picea abies) mixed with European larch (Larix decidua) and silver fir (Abies alba) T2.03TLS data and analysis Demostration area Overview of the Demo area
- 24. Kick-off Meeting 8-9/jan/2014 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). T2.03TLS data and analysis Spatial analysis
- 25. Kick-off Meeting 8-9/jan/2014 UAV Photo UAV image Source UAV Resolution 0.1 m Bands RGB-Nir T2.03TLS data and analysis Spatial analysis
- 26. Kick-off Meeting 8-9/jan/2014 DTM 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. T2.03TLS data and analysis Spatial analysis UAV DSM
- 27. Kick-off Meeting 8-9/jan/2014 Digital Height Model T2.03TLS data and analysis Spatial analysis Similar type of trees and tree height was represented with differences up to 20m. 22m aprox. 45 m aprox.
- 28. Kick-off Meeting 8-9/jan/2014 UAV Digital Surface Model T2.03TLS data and analysis Spatial analysis •In very steep terrain additional ground points are required •Digital Surface model was not used in this forest.
- 29. Kick-off Meeting 8-9/jan/2014 Tree Counting T2.03TLS data and analysis Spatial analysis
- 30. Kick-off Meeting 8-9/jan/2014 Inventory Planning is Essential in order to: 1.Ensure efficient and effective data collection 2.Minimise cost 3.Optimise accuracy T2.03TLS data and analysis Inventory plan Where the samples must be taken? •Near cable crane •All species trees types must be represented (stratification) •Inventory costs and accessibility •Variability of the forest stand •Target accuracy •Number of tree market in the TLS plot How many samples? Cable crane
- 31. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Inventory plan
- 32. Kick-off Meeting 8-9/jan/2014 The advantage of in field application: -2 way communication between office and field using GPRS or Wifi -Pre-loaded sample area location - Easy navigation in field for operator - Bluetooth recording with callipers - Stem defects recording T2.03TLS data and analysis Field measurement
- 33. Kick-off Meeting 8-9/jan/2014 TLS Height recording hypsometer DBH recording T2.03TLS data and analysis Field measurement
- 34. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Field Measurement
- 35. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Field Measurement
- 36. Kick-off Meeting 8-9/jan/2014 Tree detection and branches removal (Autostem) T2.03TLS data and analysis TLS analysis
- 37. Kick-off Meeting 8-9/jan/2014 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 •Stem defects (Straightness) T2.03TLS data and analysis TLS analysis
- 38. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Data inference Tree level estimation High res. image Tree level regression Tree detection Tree detection TLS samples 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) Regression analysis
- 39. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Inventory plan y = 15.595Ln(x) - 3.3124 R2 = 0.5709 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 Tree Crown (m2) DBH (cm) y = 0.2277x + 23.766 R2 = 0.3594 0 5 10 15 20 25 30 35 40 45 0 10 20 30 40 50 60 70 Tree Crown (m2) Height (m) Only the tree crown size can be used for the estimation of the tree parameters
- 40. Kick-off Meeting 8-9/jan/2014 Predicted DBH= F(crown size) Predicted Height = F(DBH) Predicted Volume = Taper equations (DBH, height) T2.03TLS data and analysis Spatial analysis y = -0.0023x2 + 0.5775x + 8.3554 R2 = 0.7706 0 5 10 15 20 25 30 35 40 45 0 20 40 60 80 100 Height (m) DBH (cm)
- 41. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Spatial analysis
- 42. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Known limitations 1.- Tree level inference •High resolution (lower than 30 cm) required for tree detection (e.g. UAV) •Hidden trees under forest canopy can be height (63%) in mixed stand. Ideally to be used in even-age plantations •Field tree locations (GPS) do not match correctly tree crown detected by UAV (Ideally to be used in area with good GPS signal and use corrections when required) •Limited species detection (ideally to be used in pure stands) •Relative good correlation between Volume and Crown/Height (around 70%)
- 43. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Implementation AutoStemTM TLS Cutting simulator SLOPE Field App TM DB Manual measurement Field Terminal SD card Bluetooth Internet Field Site Treemetrics Service Centre Update Slope DB SLOPE DB SLOPE Service Centre Treemetrics Intranet TM Process Control Forest Information System Office Site •Information display •Reports •Updates Update API
- 44. Kick-off Meeting 8-9/jan/2014 T2.03TLS data and analysis Implementation •The use of TLS in forest inventory improves significantly the estimation of tree volume •The use of UAV image (0.2cm) or medium resolution EO (0.5m) can improve significantly the field inventory planning. •The combination of TLS and UAV image can improve significantly the estimation and location of the timber in the forest. •The analysis of allometric relationships at tree level has result successful using UAV stereo analysis and very limited using EO data . •The estimation of the timber in the forest need to take into the trees hidden trees