Favicon of Cognex

Bosch

Bosch increases error catch rate to 99-100% with machine vision at Czech plant

99-100%Error Catch Rate (After)
85%Error Catch Rate (Before)
75Wireless ID Readers Deployed

The Challenge

Bosch Diesel s.r.o. in Jihlava, Czech Republic manufactures high-precision automotive components for the Common Rail diesel injection system — including high-pressure injection pumps, rails, and pressure control valves — where permitted deviations are measured in micrometers. At that tolerance level, an 85% error catch rate was operationally unacceptable: defective pump bodies were reaching downstream stations and risking shipment to customers. Beyond accuracy, the manual scanning process created workplace hazards, as operators carried wired readers across slippery production floors. The plant also lacked reliable traceability, with no consistent way to track components through all key operations using Data Matrix codes.

The Solution

Bosch deployed approximately 80 Cognex ID readers — roughly 75 of them wireless — alongside In-Sight machine vision systems across twelve sections of the CP3 department. The Cognex DataMan readers verify the legibility of 6mm Data Matrix codes imprinted on pump bodies, while In-Sight vision systems control work and finishing operations on production lines, detect defects, and identify error sources in real time. Fixed-mount barcode readers replaced handheld manual scanning at key stations. Configuration was handled through the Spreadsheet programming interface in Cognex's In-Sight Explorer software, with initial training delivered by Cognex and subsequently managed internally. Where ambient lighting was insufficient, infrared external lighting was integrated and controlled via PLC alongside cameras and blowers used to clear debris from DMC surfaces before reading.

Results

The headline outcome was a near-elimination of missed defects: error catch rate rose from 85% to 99–100%, directly reducing the distribution of defective components to downstream stations. Supporting that improvement:

  • 75 of ~80 ID readers deployed wirelessly, eliminating cabling hazards on slippery floors
  • Processing times reduced and labor costs decreased through fixed-mount automation
  • Full component traceability established — operation time, worker name, machine, spindle, DMC code quality, and body status all recorded per part
  • Back-search capability enabled via recorded production images, supporting root-cause analysis

The ease of use and reliability of the system prompted Bosch Jihlava's Diesel Systems division to plan the purchase of three additional vision systems.

Key Takeaways

  • Wireless deployment removes a hidden barrier — in wet or slippery industrial environments, eliminating cable management is as important as the vision technology itself.
  • DMC traceability multiplies the value of inspection — recording per-part metadata (operator, machine, spindle, timestamp) transforms defect detection into a root-cause tool.
  • Environmental factors must be engineered in — ambient lighting variability and surface contamination (dirt, coolant) require purpose-built solutions like infrared lighting and air blowers, not afterthoughts.
  • Internal training ownership sustains adoption — Bosch transitioned from vendor-led to internally managed training, reducing long-term dependency and accelerating operator onboarding.
  • Proven ROI drives organic expansion — a well-scoped initial deployment that demonstrates reliability is the most effective justification for expanding to additional lines.

Share:

Vendor

Favicon of CognexCognex

Details

Industry
Automotive
AI Technology
Computer Vision
Company Size
Enterprise
Company
Bosch
Quality
Verified

Have a similar implementation?

Share your customer's AI results and link it to your vendor profile.

Submit a case study →