Evaluation of Unmixing Methods for Impervious Surface Area Extraction From Simulated EnMAP Imagery
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Distribution of impervious surface area (ISA) is an important input in a wide range of urban ecosystem studies. The future launch of the German hyperspectral satellite environmental mapping and analysis program (EnMAP) in 2019 provides new opportunities for timely and global ISA extraction. The previously proposed EnMAP applications heavily relied on existing reference endmembers, which may be impractical on a global scale. To overcome this defect, we suggest to use the nonnegative matrix factorization (NMF) method to extract the endmember directly from EnMAP imagery. Three traditional unmixing method (e.g., N-Findr, pixel purity index, and independent component analysis) and four NMF-based methods with different constraints (e.g., sparseness, convex volume, and nonlinearity) were used to obtain series of endmember sets, ISA fraction, and classification maps. In results, the NMF-based methods outperformed the three traditional unmixing method, by achieving 0.5-0.6 R-squared values in the linear regression models between predicted and reference ISA percentages, and over 85% overall accuracy in ISA classification maps. We found that the NMF-based spectral unmixing methods are suitable to work with the EnMAP image, when reference endmember data are unavailable. In addition, we processed the widely used Hydice urban test image with the same methods and compared the resulting ISA percentage/classification maps with the EnMAP results, considering the different features of the Hydice and EnMAP sensors. In the results, it is proved that the EnMAP image has great potential in ISA mapping on a global scale, with reasonable overall accuracy and economical efficiency.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it