Classification of Compact Polarimetric Synthetic Aperture Radar Images
Bibliographic record
Abstract
The RADARSAT Constellation Mission (RCM) was launched in June 2019. RCM, in addition to dual-polarization (DP) and fully quad-polarimetric (QP) imaging modes, provides compact polarimetric (CP) mode data. A CP synthetic aperture radar (SAR) is a coherent DP system in which a single circular polarization is transmitted followed by the reception in two orthogonal linear polarizations. A CP SAR fully characterizes the backscattered field using the Stokes parameters, or equivalently, the complex coherence matrix. This is the main advantage of a CP SAR over the traditional (non-coherent) DP SAR. Therefore, designing scene segmentation and classification methods using CP complex coherence matrix data is advocated in this thesis. \n \nScene classification of remotely captured images is an important task in monitoring the Earth's surface. The high-resolution RCM CP SAR data can be used for land cover classification as well as sea-ice mapping. Mapping sea ice formed in ocean bodies is important for ship navigation and climate change modeling. The Canadian Ice Service (CIS) has expert ice analysts who manually generate sea-ice maps of Arctic areas on a daily basis. An automated sea-ice mapping process that can provide detailed yet reliable maps of ice types and water is desirable for CIS. In addition to linear DP SAR data in ScanSAR mode (500km), RCM wide-swath CP data (350km) can also be used in operational sea-ice mapping of the vast expanses in the Arctic areas. The smaller swath coverage of QP SAR data (50km) is the reason why the use of QP SAR data is limited for sea-ice mapping. \n \nThis thesis involves the design and development of CP classification methods that consist of two steps: an unsupervised segmentation of CP data to identify homogeneous regions (superpixels) and a labeling step where a ground truth label is assigned to each super-pixel. An unsupervised segmentation algorithm is developed based on the existing Iterative Region Growing using Semantics (IRGS) for CP data and is called CP-IRGS. The constituents of feature model and spatial context model energy terms in CP-IRGS are developed based on the statistical properties of CP complex coherence matrix data. The superpixels generated by CP-IRGS are then used in a graph-based labeling method that incorporates the global spatial correlation among super-pixels in CP data. \n \nThe classifications of sea-ice and land cover types using test scenes indicate that (a) CP scenes provide improved sea-ice classification than the linear DP scenes, (b) CP-IRGS performs more accurate segmentation than that using only CP channel intensity images, and (c) using global spatial information (provided by a graph-based labeling approach) provides an improvement in classification accuracy values over methods that do not exploit global spatial correlation.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".