The purpose of treespectra code is to take ground or image-based tree reflectance spectra, build a classifier or regression ,and apply that model to an imaging spectrometer data cube (aka hyperspectral image).
1) First, install dependencies /Scripts/install_dependencies.R 2) Then source the package of lecospectR /Functions/lecospectR.R
The functions in lecospectR are divided into sections, all of which are loaded by sourcing /Functions/lecospectR.R
/Functions/dataframe_operations.R
/Functions/model_support.R
/Functions/pfts.R
/Functions/pipeline.R
/Functions/raster_operations.R
/Functions/spectral_operations.R
/Functions/training_utilities.R
/Functions/type_conversion.R
/Functions/utilities.R
/Functions/validation.R
1) Ground spectra were collected around Maine.
2) Locations where ground spectra were collected
1) Our workflow assumes a list of tree species from a table. This table is used for several steps, including aggregating validation to the same taxonomic level as training data. 2) Build a spectral library from a range of field scans collected with Tungsten halogen illumination using a leaf clip or contact probe. These scripts delete bad scans and standardize the associated information into a single metadata format. Around 90 vegetation indices are also calculated and the narrow band reflectance is resampled and smoothed to 5 nm bands.
/Scripts/2_DataMunging.R
/Scripts/2B_DataMunging_missing_spectra.R
/Scripts/3_Create_SpecLibPSR.R
The output of running these scripts are two spectral libraries, one of reflectance and one of vegetation indices along with the metadata for these reflectance measurements.
/Output/C_001_SC3_Cleaned_SpectralLib.csv
/Data/D_002_SpecLib_Derivs.csv
3) Collect spectra from pixels in UAV-captured images of tree canopies that are visible in imagery. Stem maps of known tree locations are overlaid on UAV flight path extents to identify canopy sources of spectra (1_MSGC_extent, 2_MSGC_get_trees). Reflectance spectra are extracted from hand-digitized tree canopies provided as a shapefile to the script (3_crop_image_canopy_ROI). Each canopy is associated with three hand-digitized shapefiles: 1) full canopy, 2) illuminated-only branches, and 3) shaded-only branches. Cropping each image by the shapefile creates a set of images with 326 bands from 400-1000nm, with each image representing the spectral profile of one tree canopy. Script (4_Parse_ROI_canopy_spectra) adds metadata to each pixel by tree species and outputs a spectral library in wide format. Scripts (5_Clean_image_spectra) and (6_Preprocess_viz_image) clean and process data to facilitate the visualization of reflectance data by species.
/Scripts/2_Image_processing/Headwall/1_MSGC_extent.R
/Scripts/2_Image_processing/Headwall/2_MSGC_get_trees.R
/Scripts/2_Image_processing/Headwall/3_Crop_image_canopy_ROI.R
/Scripts/2_Image_processing/Headwall/4_Parse_ROI_canopy_spectra.R
/Scripts/2_Image_processing/Headwall/5_Clean_image_spectra.R
/Scripts/2_Image_processing/Headwall6_Preprocess_viz_image.R
After running these scripts, the outputs include the relectance for each pixel from the canopies digitized from images. These data can be found in the MSGC_DATA directory (PEF example is shown here).
M:/MSGC_DATA/PEF-Demerit/Spectral_libraries/
After cleaning and pre-processing, the reflectance data can be summarized and visualized in various ways. Below is a summary of the median (purple) and interquartile ranges (75% dark grey & 95% light grey) of reflectance for a number of forest tree species in Maine. Median reflectance of fully-illuminated branches (orange) and fully-shaded branches (blue) are also shown. Sample size in number of individuals and number of spectral pixels is displayed for each species.
All UAV flights for Maine are shown in the map in the first image below (top left). Flight extents are generated from /Scripts/2_Image_processing/Headwall/1_MSGC_extent. These locations are then collected and summarized using /Scripts/2_Image_processing/Headwall/7_Visualize_spectra_locations, which produces the flight locations plotted on the map. The second image below (top right) shows all flight extents at the Penobscot Experimental Forest. Each overlapping rectangle polygon represents a single hyperspectral image (aka imaging spectrometer data cube). The third image below (bottom left) shows a screen capture of one data cube. The fourth image below (bottom right) shows a portion of the same data cube enlarged, with hand-digitized shapefiles of tree canopies overlaid.
Currently in development