ToolsRTM
  • Home
  • R Tutorials
  • Python Tutorials
  • ← Back to RTM-Suite

R Tutorials

TOOLSRTM · R PKGDOWN ARTICLES

R Tutorials

Nineteen rendered R vignettes, grouped by theme: the models themselves, trait inversion (classical ML and deep learning), sensitivity analysis, and satellite/spatial applications. Every number and figure is real output from an actual run.

Leaf, canopy, soil & atmosphere models · t01–t09, t16
Five leaf models × three canopy models × two soil models × SPART for the full soil-canopy-atmosphere chain.
01

Getting Started with ToolsRTM

A first simulation, start to finish.

02

From Leaf to Canopy Reflectance

Chaining a leaf model into a canopy model.

03

SPART: Soil-Plant-Atmosphere RT

Top-of-canopy vs. top-of-atmosphere reflectance.

04

Comparing Radiative Transfer Models

All 15 leaf×canopy combinations, side by side.

05

Building Look-Up Tables

Parameter sampling for a simulation LUT.

06

Large-Scale and Parallel RTM Simulation

Running a LUT across many cores.

07

Sensor Convolution

Native spectrum → real sensor bands.

08

Hyperspectral and VNIR Sensor Convolution

PRISMA and other hyperspectral sensors.

09

Vegetation Indices and Spectral Features

NDVI, red-edge and friends, from simulated bands.

16

MARMIT + fourSAIL + SPART

Realistic, moisture-dependent soil reflectance inside a full canopy run.

Trait inversion & machine learning · t11–t14
Going backwards: from a reflectance spectrum to the vegetation trait that produced it.
11

From Physics to Vegetation Traits: Hybrid Inversion

LUT-trained models applied to real and synthetic spectra.

12

Comparing ML Algorithms for RTM Inversion

Random Forest, SVM, PLSR and ensembles, head-to-head.

13

Deep Learning for RTM Inversion

Dense networks and a 1D-CNN on hyperspectral bands, vs. Random Forest.

14

End-to-End RTM Inversion Pipeline

Simulate → convolve → invert, as one pipeline.

Sensitivity analysis · t10
Which trait actually moves the signal, and where in the spectrum.
10

Understanding RTM Sensitivity

One-at-a-time sweeps, Sobol indices, and Johnson relative weights.

Satellite & spatial applications · t15, t17–t19
Real pixels: actual Sentinel-2 scenes via STAC, pixel-by-pixel inversion, and FLEX sub-pixel heterogeneity.
15

From Satellite Reflectance to Traits: Real EO Application

Inverting a real, not simulated, scene.

17

Monitoring a Forest Site Through Time

A real forest site, tracked across a time series.

18

Data-Driven Spatial Index Mapping

Trait maps from real Sentinel-2 imagery.

19

FLEX Cal/Val: ESU Heterogeneity Mapping

Scaling up to FLEX's coarser footprint with sub-pixel heterogeneity.

The full package reference (every function, every argument) lives at the ToolsRTM pkgdown site; the complete numbered tutorial index is on the RTM-Suite tutorials overview.

RTM-Suite · ToolsRTM tutorials & docs

 

R and Python kept numerically in sync