refactor: render Q10 overlays with V1

[tubededentifrice]

Summary

Second rendering slice extracted from #900, now rebased onto main after #902 merged. Q10 overlays reuse the V1 image generator for paths, robot/dock markers, zones, and virtual walls while Q10 parsing remains responsible for converting protocol coordinates into MapData.

Changes

• keep the 2.5 mm trace grid separate from the 5 mm vector calibration
• derive idle erase/restriction state from the map header
• use the saved dock position and heading from absolute header coordinates
• draw paths, robot/dock markers, no-go/mop zones, and virtual walls through the shared V1 renderer
• keep renderer construction private and keep this slice free of trait, transport, and CLI changes

Before / after

The same isolated MapData fixture is rendered before and after this slice.

Before	After

Validation

• pytest -q — 887 passed, 86 snapshots passed
• Ruff lint and format checks — all passed
• git diff --check — passed

Stack / review order

1. feat: align Q10 map colors with V1 #902 — base raster and room palette (merged)
2. This PR: V1 overlay renderer
3. fix: show docked Q10 robot on map #904 — docked-state composition

This branch is rebased onto main, so Files changed contains only the overlay-renderer slice. The transport correction is independent in #901; #900 remains the aggregate real-device/Home Assistant reference.