Bill of Materials Recognition
Bill of materials recognition (BOM extraction) – extract BOMs from PDF
This tool extracts bills of materials (BOM) from technical PDF documents and detects item numbers, quantities, article numbers and other structured data from tables and drawing documents.
The PDF file is processed server-side and automatically deleted after analysis.
What is BOM recognition used for?
Bills of materials are a key basis for design, purchasing, manufacturing, ERP and PDM systems. The tool automates data capture from technical PDF documents.
Typical applications:
- transfer of BOMs into ERP systems
- import into PDM and PLM systems
- review of supplier documents
- digitisation of archive documents
- preparation of manufacturing data
- comparison of items and quantities
Which data is recognised?
Depending on the document and table structure, various information can be recognised automatically.
- item numbers (Pos.)
- quantities (pcs, qty)
- article numbers
- part numbers
- part designations
- material data
- further table information
Supported document types
- PDF files with embedded tables
- technical drawings with BOMs
- manufacturing drawings
- assembly drawings
- scanned PDFs with OCR support
How does BOM recognition work?
- Upload a PDF file
- Analyse tables and content
- Detect columns automatically
- Assign items, quantities and article numbers
- Generate a structured BOM
- Display or export the results
Technologies used
The analysis uses proven open-source libraries:
- pdfplumber
- Camelot (optional)
- OCR technologies for scanned documents
- Python-based document analysis
Typical benefits
- time saved in data capture
- fewer manual input errors
- fast digitisation of existing documents
- easy further processing in ERP and PDM systems
- automatic evaluation of technical documents
Current limitations
- Current maximum file size is 80 MB per PDF file
- By default the first pages are analysed
- Scanned PDFs without a clear table structure yield fewer hits
- Recognition quality depends on document quality and layout
- Extended table recognition may require additional system components