LiDAR based Forest Inventory: Development and Validation of Automated DBH extraction algorithm for Pacific Northwest Douglas Fir Forest.
Graduate Student Name
Avishek Hamal
Email Address
Faculty mentor/Supervisor
Dr. Bogdan Strimbu
Email Address (Faculty mentor/Supervisor)
Department Affiliation
Forest Engineering Resources & Management
Job Location
Peavy Forest Science Center, with occasional field visit to Elliott State Research Forest
Description of project or research opportunity
The major portion of this project involves the concept of point cloud geometry, and analysis, which will be handled by graduate student. The intended workload for the undergraduate student involves the concept of Forest Mensuration, Forest Measurement, Geographic Information System and Coordinate Systems, Cartography and Map Making, Basics of Point Cloud Georeferencing and Measurement.
Project Background and Goals: Recent developments in forest inventory technology such as LiDAR have been a growing area of research due to its ability to collect fully digitalized versions of a forest plot and centimeter level tree measurement accuracy obtained from the point cloud. It also offers tremendous efficiency in terms of time required in collection of data: while a manual measurement requires 20-45 minutes, a modern handheld LiDAR such as FJD Trion P1 Scanner can do the job in less than 10 minutes. Researches have been done in LiDAR based forest inventory in urban forest, and plantation forests, however, much research work is required to experiment and validate the performance of these modern handheld LiDAR in complex forest of Pacific Northwest (PNW) where the undergrowth might introduce noise in the data and huge trees growing up to 200cm in diameter and almost 100m in height create challenges of shadowing and ranging limitations for the scanners. Therefore, this project is intended to devise an automated algorithm to extract tree DBH measurement from point cloud and evaluate it in the context of complex PNW forests.
Project Background and Goals: Recent developments in forest inventory technology such as LiDAR have been a growing area of research due to its ability to collect fully digitalized versions of a forest plot and centimeter level tree measurement accuracy obtained from the point cloud. It also offers tremendous efficiency in terms of time required in collection of data: while a manual measurement requires 20-45 minutes, a modern handheld LiDAR such as FJD Trion P1 Scanner can do the job in less than 10 minutes. Researches have been done in LiDAR based forest inventory in urban forest, and plantation forests, however, much research work is required to experiment and validate the performance of these modern handheld LiDAR in complex forest of Pacific Northwest (PNW) where the undergrowth might introduce noise in the data and huge trees growing up to 200cm in diameter and almost 100m in height create challenges of shadowing and ranging limitations for the scanners. Therefore, this project is intended to devise an automated algorithm to extract tree DBH measurement from point cloud and evaluate it in the context of complex PNW forests.
Tasks student will perform
-Collect forest cruising data.
-Navigate the inventory plot location coordinates loaded in GPS
-Should understand inventory data collection process using point sampling (variable radius plot)
-Use a Diameter tape to measure the DBH
-Use a rangefinder to measure the distance of tree stem from the plot center
-Operate an FJD Trion P1 and FJD Trion S1 Scanners for LiDAR scan of the measured plot
-Register and pre-process the point cloud to remove noises and flag erroneous data.
-Generate terrain models/canopy models from point cloud.
-Work with canopy models in ArcGIS/QGIS.
-Be able to read scientific papers and summarize the findings (preferred but not required).
-Have a driver’s license.
-Navigate the inventory plot location coordinates loaded in GPS
-Should understand inventory data collection process using point sampling (variable radius plot)
-Use a Diameter tape to measure the DBH
-Use a rangefinder to measure the distance of tree stem from the plot center
-Operate an FJD Trion P1 and FJD Trion S1 Scanners for LiDAR scan of the measured plot
-Register and pre-process the point cloud to remove noises and flag erroneous data.
-Generate terrain models/canopy models from point cloud.
-Work with canopy models in ArcGIS/QGIS.
-Be able to read scientific papers and summarize the findings (preferred but not required).
-Have a driver’s license.
Special skills required
-Basic Understanding of Georeferencing and Coordinate System
-Working Knowledge of GIS.
-Understanding and basic working knowledge of point clouds and point cloud analysis software: QT Modeler, CloudCompare. This skill can be learned, but having prior understanding is preferred.
-Field knowledge of Forest Inventory process.
-Be able to distinguish major Pacific Northwest Forest Species: Douglas Fir, Western Hemlock, Grand Fir, Western redcedar, Bigleaf Maple and such.
-Reading Scientific Articles and Summarizing them (preferred but not required).
-Working Knowledge of GIS.
-Understanding and basic working knowledge of point clouds and point cloud analysis software: QT Modeler, CloudCompare. This skill can be learned, but having prior understanding is preferred.
-Field knowledge of Forest Inventory process.
-Be able to distinguish major Pacific Northwest Forest Species: Douglas Fir, Western Hemlock, Grand Fir, Western redcedar, Bigleaf Maple and such.
-Reading Scientific Articles and Summarizing them (preferred but not required).