Record accuracy sweep results for corridor

[brendancol]

Summary

• Audited xrspatial/corridor.py against the 5 accuracy categories (precision loss, NaN/Inf propagation, off-by-one, projection corrections, backend inconsistency).
• No CRIT/HIGH issues. One LOW: corridor inherits float32 from cost_distance so normalized = corridor - corridor_min can lose precision near the min when costs accumulate to large values. The arithmetic in corridor itself does not introduce extra loss.
• NaN handling is correct: corridor.min() uses skipna=True, the result is checked with np.isfinite before normalization, and unreachable corridors return all-NaN.
• All four backends route through pure xarray arithmetic. The threshold path dispatches to dask.where, cupy.where, or np.where via a try/except pattern.

Test plan

• CSV updated with module, date, severity, categories, notes

[brendancol]

@copilot resolve the merge conflicts in this pull request

[Copilot]

@copilot resolve the merge conflicts in this pull request

Merge conflicts resolved in 30c60be. The CSV now includes the corridor row from this branch alongside all rows added to main since the branch was created (bilateral, contour, edge_detection, diffusion, glcm, mahalanobis, normalize, sky_view_factor, worley).