specimen-aware evaluation of hyperspectral recognition of construction and demolition waste

Python code and result files for the manuscript "Beyond random window splits: specimen-aware evaluation of hyperspectral recognition of construction and demolition waste" by U.G.U.P. Samarasekara, S.A.R.R. Suraweeraarachchi, J.A.S.C. Jayasinghe, U. Jayasinghe and C.S. Bandara,. The study re-analyses two public visible to near-infrared hyperspectral datasets of construction and demolition waste, treating the physical piece rather than the pixel or window as the unit of replication. It asks where the spectral information lies once differences between pieces are accounted for (Q1), how much a reported accuracy depends on the way the data are split (Q2), how well performance holds up at the decisions a sorting plant makes (Q3), and whether trained models carry over between the two datasets (Q4). Files- code.zip: analysis scripts (Python 3.11 or later), list of required packages and the MIT licence.- results.zip: harmonised window tables, mask previews and every result file behind the figures and tables.- README_code.pdf and README_results.pdf: instructions and file descriptions. Unzip both archives into the same folder. The figures and tables can be rebuilt from the stored results without the raw data, which are available from the original records: Calcagni et al. (2025), https://doi.org/10.5281/zenodo.15976821, and Vítek et al. (2024), https://doi.org/10.5281/zenodo.13840470.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-15
DOI
https://doi.org/10.5281/zenodo.24241603
Primary Topic
Remote-Sensing Image Classification
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

specimen-aware evaluation of hyperspectral recognition of construction and demolition waste

Upul Jayasinghe, J. A. S. C. Jayasinghe, Chaminda Senarathna Bandara, U. G. U. P. Samarasekara et al.
Zenodo (CERN European Organization for Nuclear Research)
Remote-Sensing Image Classification
article

specimen-aware evaluation of hyperspectral recognition of construction and demolition waste

Upul Jayasinghe, J. A. S. C. Jayasinghe, Chaminda Senarathna Bandara, U. G. U. P. Samarasekara, S.A.R.R. Suraweeraarachchi
article en

Abstract

Python code and result files for the manuscript "Beyond random window splits: specimen-aware evaluation of hyperspectral recognition of construction and demolition waste" by U.G.U.P. Samarasekara, S.A.R.R. Suraweeraarachchi, J.A.S.C. Jayasinghe, U. Jayasinghe and C.S. Bandara,. The study re-analyses two public visible to near-infrared hyperspectral datasets of construction and demolition waste, treating the physical piece rather than the pixel or window as the unit of replication. It asks where the spectral information lies once differences between pieces are accounted for (Q1), how much a reported accuracy depends on the way the data are split (Q2), how well performance holds up at the decisions a sorting plant makes (Q3), and whether trained models carry over between the two datasets (Q4). Files- code.zip: analysis scripts (Python 3.11 or later), list of required packages and the MIT licence.- results.zip: harmonised window tables, mask previews and every result file behind the figures and tables.- README_code.pdf and README_results.pdf: instructions and file descriptions. Unzip both archives into the same folder. The figures and tables can be rebuilt from the stored results without the raw data, which are available from the original records: Calcagni et al. (2025), https://doi.org/10.5281/zenodo.15976821, and Vítek et al. (2024), https://doi.org/10.5281/zenodo.13840470.

Zenodo (CERN European Organization for Nuclear Research)
Remote-Sensing Image Classification
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon