Getting Started with cuvis.r • cuvis.r

Overview

cuvis.r provides R bindings to the Cubert CUVIS C SDK for reading, processing, and exporting hyperspectral imaging data from Cubert snapshot cameras. It mirrors the API of cuvis.python.

Prerequisites

Install the Cubert CUVIS SDK (>= 3.4.0) from https://cloud.cubert-gmbh.de/s/qpxkyWkycrmBK9m
Set the CUVIS_SDK environment variable to the SDK install directory
Install cuvis.r from GitHub:

remotes::install_github("cttir/cuvis.r")

Quick Start

Initialize the SDK

Every cuvis.r session begins with cuvis_init() and ends with cuvis_shutdown():

library(cuvis.r)

cuvis_init()

Load a Session File

Cubert cameras store measurements in .cu3s session files. Load one with cuvis_session():

session <- cuvis_session("path/to/measurement.cu3s")
print(session)

Extract a Measurement

Session files contain one or more measurements. Access them by index (1-based):

mesu <- cuvis_get_measurement(session, 1)
print(mesu)

Inspect Metadata

md <- cuvis_get_metadata(mesu)
md$name
md$integration_time
md$serial_number
md$product_name

Get the Data Cube

The hyperspectral data cube is returned as a 3D R array [rows, cols, bands] with wavelength metadata attached:

# First, process to generate the cube
ctx <- cuvis_processing_context(session)
cuvis_reprocess(ctx, mesu, mode = "raw")

cube <- cuvis_get_cube(mesu)
dim(cube)                        # e.g., c(100, 100, 61)
attr(cube, "wavelengths")        # wavelengths in nm

You can subset bands, compute means, or use standard R array operations:

# Mean reflectance per band
band_means <- apply(cube, 3, mean, na.rm = TRUE)

# Extract a single band (e.g., band 30)
single_band <- cube[, , 30]
image(single_band, main = "Band 30")

Export to Standard Formats

# ENVI format (widely supported by remote sensing tools)
cuvis_export_envi(mesu, "output/envi/")

# Multi-channel TIFF
cuvis_export_tiff(mesu, "output/tiff/")

Shutdown

cuvis_shutdown()

Next Steps

See vignette("cuvis-r-calibration") for the full calibration workflow (dark/white references, reflectance conversion)
See the cuvis.python examples for workflow patterns that translate directly to cuvis.r

Use of LLM tools

Portions of this package were prepared with assistance from large language model tooling for narrowly defined, non-authorial tasks: copyediting, prose smoothing, Markdown/LaTeX formatting, scaffolding of boilerplate files (CI configs, build scripts), code refactoring. The tools used were Chat AI, the LLM service of KISSKI (GWDG), and a self-hosted Mistral Small (24B, Apache-2.0) run locally via Ollama and the ollamar R package — local inference only, with no data sent to third parties for the self-hosted model.

All scientific claims, methodological choices, analyses, interpretations, and conclusions are the author’s own. No LLM-generated text was incorporated without review and revision, and every reference was verified against its DOI, arXiv ID, or ISBN.