Geospatial Forest Analysis

Course Description

Review of passive and active sensors used in remote sensing. Geospatial data manipulation in R. Image classification. Point clouds: lidar and photogrammetric. Point cloud manipulation for forestry application. Sampling for accuracy assessment. Lec/lab. PREREQS: FE 444 or GEOG 480 or equivalent.

Learning Resources

All required learning resources are posted online on Canvas. However, the course requires understanding of the remote sensing techniques. Therefore, two textbook are recommended for the successful completion of this class:

  1. John Jensen (2018) Introductory Digital Image Processing 4Th Edition, Pearson
  2. Dong P. and Chen Q. (2018). Lidar remote sensing and applications . Taylor and Francis

Supplementary material: Lillesand, T.M., Kiefer, R.W. and Chipman, J (2015) Remote sensing and Image Interpretation (7th Ed). Wiley and Sons, New York.

Measurable Student Learning Outcomes

Students completing this course will acquire advanced knowledge of the techniques available extract detailed information about vegetation from remotely sense data. The course has two objectives:

  1. Proficiency in processing data in R
  2. Extraction of forestry relevant information from multispectral images and point clouds.

Course success is measured as 70 % or higher average for labs, midterm, and final  exam.

Measureable learning outcomes:

  • Students completing this course will be able to conceptualize scientific problems and break them into programmable steps (flow chart)
  • Students completing this course will be able to manipulate geospatial data in R, such as registration, co-registration, or coordinate conversion
  • Students completing this course will be able to classify and interpret images classified with spatial, spectral, and combined algorithms to answer forest relevant questions
  • Students completing this course will be able to extract information from point clouds, particularly ground, crowns, and estimate trees and stands heights.
  • Students completing this course will be able to design a sampling strategy  for assessing the performances of the image classification algorithms

Course Content

Week

Topic

Lab

1

Introduction to R

 

Lab 1: Gdal and raster packages

Input/output data, Coordinate systems

2

Remote sensing data

Lab 2: manipulation of images and points clouds

Registration and co-registration

3

Image classification I

Lab 3: supervised and unsupervised classification

ISODAT, ML, SAM, & Confusion matrices

4

Image classification II

Lab 4: advanced image classification techniques

Caret & Random forests

5

Point clouds

Lab 4: advanced image classification techniques (2)

Bayesian Models & Support Vector Machine

6

Point clouds

Lab 6: lidar and photogrammetric point clouds

lidR, pdal, & DTM extraction

7

Forest measurements

Lab 7: tree and stand attributes

CHM, segmentation of individual trees

8

Assessment of results accuracy

Lab8: sampling for accuracy assessment

9

Project topics

Lab9: Individual / Group work on project

10

 

Lab 10: Individual / Group work on project

Finals

 

Final Exam