COURSES
ForestSat 2016 is glad to host several courses related to the use of remote sensing applied to forestry.
Classrooms are organized inside the conference building and are finalized to work on real data. For this reason participants need to bring their own laptops.
The cost of each course is US$30 per person. The participation to the courses is strictly restricted to regularly registered ForestSat 2016 participants. For register in the courses, you must fill the Course Registration Form from your user account on the website.
All courses will be conducted after the conference, on Friday 18th November in the morning.
Course #1: Forest Canopy Height Mapping with Radar Interferometry
Instructors: Naiara Pinto, Marc Simard, Michael Denbina.
Maximum number of participants: 30
Language: English
Description:
Forest canopy height is an essential biophysical parameter for monitoring forest ecosystems. Under specific observation scenarios, Synthetic Aperture Radar sensors can be used to produce wall-to-wall canopy height maps through a technique called Polarimetric Interferometric Synthetic Aperture Radar (PolInSAR). Data amenable for PolInSAR have been collected by US American and European airborne sensors, and are expected to be available from ESA’s BIOMASS radar mission, planned to launch in 2020.
This workshop presents a first introduction of the PolInSAR technique. Students will learn basic SAR concepts and work on a hands-on activity with simulated SAR data. We will cover the following topics:
- Typical scattering mechanisms at forest sites.
- PolInSAR sensor observation strategy.
- Canopy height estimation using Random Volume over Ground model.
- Uncertainty sources: temporal decorrelation and spatial heterogeneity.
- Examples from temperate, tropical, and boreal forest.
- Synergies with LiDAR-derived metrics.
Course #2: LiDAR
Instructor: Martin Isenburg.
Maximum number of participants: 50
Language: English
Description:
The LiDAR course will focus on how to create Canopy Height Models and plot as well as raster based forestry metrics such as canopy cover, as well as percentiles, strata, kurtosis, skewness, standard deviation, etc for both normalized heights as well as intensities. The focus of the course would be the many important choices that need to be made and their impact on the resulting output. The course will also emphasize the importance of quality checking the raw LiDAR obtained before starting to produce the needed outputs.
(1) Quality Checking
- Validating LAS/LAZ files for ASPRS specification conformance to assure – in particular – the correctness of return numbering as this affects pulse density, canopy cover, rumple, and any other return based metric.
- Assuring presence of flightline information and possible ways to reconstruct it if missing.
- Using flightline information for checking proper alignment and overlap of flight lines.
- Assuring the consistency of intensity values across flightlines and – if needed – how to correct them before computing intensity metrics.
- Detecting the presence of clouds and their impact on intensity, point density, and return-based metrics.
(2) Preprocessing LiDAR
- Tile-based processing and importance of maintaining buffers around tiles to avoid edge artifacts.
- Removal of isolated low and high noise points and clouds.
- Ground classification.
- Height-normalization.
- Possible impact of height-normalization on tree crown shape and treetop detection.
- Automated detection of buildings and vegetation in rural or urban terrains.
- Efficient clipping of plots from a large LiDAR survey.
(3) Derivative Generation
- Standard DTM and DSM algorithms.
- Various approached to CHM creation. in particular how to generate a “pit-free CHM” at the highest resolution supported by the LiDAR based on the popular, award-winning method by (Khosravipour et al. 2014).
- Various approached to DSM creation and why it makes sense to use a DSM instead of a CHM to do treetop detection. time permitting we will introduce our newest algorithm (Khosravipour et al. 2015) for generating a “spike-free DSM” at the highest resolution supported by the LiDAR on the and how its benefits tree top detection.
- Generating plot-based as well as raster-based metrics for subsequent use in random-forest type algorithms for timber volume, plantation yield, or plant biomass prediction.
Course #3: Design of Forest Inventories based on Remote Sensing Data
Instructor: Günther Bronner, Umweltdata Austria.
Maximum number of participants: 50
Language: English
Description:
The proposed workshop would bring the following topics, combined with lots of examples from real projects:
- National forest inventories vs. inventories for forest companies.
- Permanent vs. temporary sample plots.
- Bitterlich angle count method vs. distinct circular sample plots.
- Linear sample plots.
- Regular vs. irregular inventory grids in plantations and natural grown forests.
- Concentration of sample plots in old grown forests.
- Stratification of forest areas by airborne LiDAR and spectral data (pre-fieldwork and post-fieldwork stratification).
- Clustering sample plots to minimize the effort of admission to sample plots.
- Forest inventory related GNSS-navigation.
- Referencing coordinates of fieldwork-assessed trees to a canopy height model.
- Combination of RS-based figures and fieldwork-based figures.
- Plausibility-check of terrestrial measured data by RS methods.
- Wall-to-wall interpolation based on RS data.
- Forest inventory related optimization of cost efficiency and accuracy.
- Representation of inventory results for operational management and decision support.
- RS-based update of inventory data after calamities.
COURSES
ForestSat 2016 is glad to host several courses related to the use of remote sensing applied to forestry.
Classrooms are organized inside the conference building and are finalized to work on real data. For this reason participants need to bring their own laptops.
The cost of each course is US$30 per person. The participation to the courses is strictly restricted to regularly registered ForestSat 2016 participants. For register in the courses, you must fill the Course Registration Form from your user account on the website.
All courses will be conducted after the conference, on Friday 18th November in the morning.
Course #1: Forest Canopy Height Mapping with Radar Interferometry
Instructors: Naiara Pinto, Marc Simard, Michael Denbina.
Maximum number of participants: 30
Language: English
Description:
Forest canopy height is an essential biophysical parameter for monitoring forest ecosystems. Under specific observation scenarios, Synthetic Aperture Radar sensors can be used to produce wall-to-wall canopy height maps through a technique called Polarimetric Interferometric Synthetic Aperture Radar (PolInSAR). Data amenable for PolInSAR have been collected by US American and European airborne sensors, and are expected to be available from ESA’s BIOMASS radar mission, planned to launch in 2020.
This workshop presents a first introduction of the PolInSAR technique. Students will learn basic SAR concepts and work on a hands-on activity with simulated SAR data. We will cover the following topics:
- Typical scattering mechanisms at forest sites.
- PolInSAR sensor observation strategy.
- Canopy height estimation using Random Volume over Ground model.
- Uncertainty sources: temporal decorrelation and spatial heterogeneity.
- Examples from temperate, tropical, and boreal forest.
- Synergies with LiDAR-derived metrics.
Course #2: LiDAR
Instructor: Martin Isenburg.
Maximum number of participants: 50
Language: English
Description:
The LiDAR course will focus on how to create Canopy Height Models and plot as well as raster based forestry metrics such as canopy cover, as well as percentiles, strata, kurtosis, skewness, standard deviation, etc for both normalized heights as well as intensities. The focus of the course would be the many important choices that need to be made and their impact on the resulting output. The course will also emphasize the importance of quality checking the raw LiDAR obtained before starting to produce the needed outputs.
(1) Quality Checking
- Validating LAS/LAZ files for ASPRS specification conformance to assure – in particular – the correctness of return numbering as this affects pulse density, canopy cover, rumple, and any other return based metric.
- Assuring presence of flightline information and possible ways to reconstruct it if missing.
- Using flightline information for checking proper alignment and overlap of flight lines.
- Assuring the consistency of intensity values across flightlines and – if needed – how to correct them before computing intensity metrics.
- Detecting the presence of clouds and their impact on intensity, point density, and return-based metrics.
(2) Preprocessing LiDAR
- Tile-based processing and importance of maintaining buffers around tiles to avoid edge artifacts.
- Removal of isolated low and high noise points and clouds.
- Ground classification.
- Height-normalization.
- Possible impact of height-normalization on tree crown shape and treetop detection.
- Automated detection of buildings and vegetation in rural or urban terrains.
- Efficient clipping of plots from a large LiDAR survey.
(3) Derivative Generation
- Standard DTM and DSM algorithms.
- Various approached to CHM creation. in particular how to generate a “pit-free CHM” at the highest resolution supported by the LiDAR based on the popular, award-winning method by (Khosravipour et al. 2014).
- Various approached to DSM creation and why it makes sense to use a DSM instead of a CHM to do treetop detection. time permitting we will introduce our newest algorithm (Khosravipour et al. 2015) for generating a “spike-free DSM” at the highest resolution supported by the LiDAR on the and how its benefits tree top detection.
- Generating plot-based as well as raster-based metrics for subsequent use in random-forest type algorithms for timber volume, plantation yield, or plant biomass prediction.
Course #3: Design of Forest Inventories based on Remote Sensing Data
Instructor: Günther Bronner, Umweltdata Austria.
Maximum number of participants: 50
Language: English
Description:
The proposed workshop would bring the following topics, combined with lots of examples from real projects:
- National forest inventories vs. inventories for forest companies.
- Permanent vs. temporary sample plots.
- Bitterlich angle count method vs. distinct circular sample plots.
- Linear sample plots.
- Regular vs. irregular inventory grids in plantations and natural grown forests.
- Concentration of sample plots in old grown forests.
- Stratification of forest areas by airborne LiDAR and spectral data (pre-fieldwork and post-fieldwork stratification).
- Clustering sample plots to minimize the effort of admission to sample plots.
- Forest inventory related GNSS-navigation.
- Referencing coordinates of fieldwork-assessed trees to a canopy height model.
- Combination of RS-based figures and fieldwork-based figures.
- Plausibility-check of terrestrial measured data by RS methods.
- Wall-to-wall interpolation based on RS data.
- Forest inventory related optimization of cost efficiency and accuracy.
- Representation of inventory results for operational management and decision support.
- RS-based update of inventory data after calamities.
0 comentarios: