How Sheet Matching Works

Summary: Bedrock’s sheet matching uses multiple signals (sheet numbers, titles, visual similarity) to automatically pair corresponding sheets between drawing revisions without manual intervention.

The Problem

When comparing drawing revisions, you need to match sheets:

  • A1.01 in the old set should compare to A1.01 in the new set
  • But what if sheet numbers changed?
  • What if sheets were added or removed?
  • What if one discipline was renumbered entirely?

Manual matching is tedious and error-prone. Bedrock automates it.

How Bedrock Matches Sheets

Signal Analysis

Bedrock analyzes multiple signals to identify matches:

SignalWeightDescription
Sheet numberHighA1.01 matches A1.01
Sheet titleMedium”First Floor Plan” matches similar titles
Visual similarityMediumOverall drawing content and layout
Title block positionLowLocation and format of title block

No single signal is definitive. The combination provides robust matching.

Matching Process

  1. Extract metadata: Read sheet numbers and titles from both sets
  2. Direct matches: Pair sheets with identical numbers
  3. Title matching: For unmatched sheets, compare titles
  4. Visual matching: For remaining sheets, analyze visual similarity
  5. Confidence scoring: Assign confidence to each match
  6. Flag ambiguous: Highlight matches needing verification

Handling Edge Cases

ScenarioHow Bedrock Handles It
Renumbered sheetsVisual similarity + title matching
New sheetsFlagged as “Added in current revision”
Deleted sheetsFlagged as “Removed from prior revision”
Split sheetsEach new sheet matched to original when possible
Combined sheetsOriginal sheets matched to combined
Different page countsUnmatched sheets listed separately

Matching Accuracy

ScenarioTypical Accuracy
Numbers unchanged99%+
Numbers changed, titles same95%+
Numbers and titles changed85-95% (may need review)
Completely reorganized setManual review recommended

When automatic matching is uncertain, Bedrock flags the sheet for human verification.

What Sheet Matching is NOT

  • Not OCR-dependent: Works even without text layer
  • Not exact matching only: Handles variations in numbering
  • Not foolproof: Ambiguous cases are flagged, not guessed
  • Not manual pairing: Automation handles the work

Verification and Override

After automatic matching:

  1. Review match summary
  2. Check flagged uncertain matches
  3. Override incorrect matches if needed
  4. Proceed to comparison

Most comparisons require no manual intervention.

Technical Details

AspectSpecification
Processing per sheetSub-second
Maximum set size1,000+ sheets
Supported formatsPDF
Language supportAny (visual matching is language-agnostic)

FAQ

What if Bedrock matches the wrong sheets?

Uncertain matches are flagged for review. You can override any match before comparison runs. Incorrect matches don’t silently proceed.

Does sheet matching work across different CAD systems?

Yes. Sheet matching analyzes the output PDF, not the source format. Drawings from AutoCAD, Revit, or any other system work the same.

What about sheets with no numbers?

Bedrock falls back to title and visual similarity. Sheets without numbers or titles are flagged for manual matching.

How long does matching take?

Matching typically completes in seconds, even for large sets. It’s part of the overall comparison processing time.

Key Takeaways

  • Sheet matching uses multiple signals (number, title, visual)
  • Automation handles standard cases without intervention
  • Edge cases (renumbering, additions, deletions) are detected
  • Uncertain matches are flagged, not guessed
  • Manual override available when needed
  • Works across any CAD system or drawing source

Last updated: 2026-02-04