Computer Vision

Photo Diagnostics of Contamination and Quality Control for Product Processing

Photo Diagnostics of Contamination and Quality Control for Product Processing

A computer vision module in which an operator photographs a product, the system estimates the level and type of contamination, helps choose the right processing mode, and then supports quality control after treatment to confirm whether the result is acceptable.

Pregled

Computer vision as an operator assistant and quality control tool

This project focused on a photo-based diagnostics system that helps production staff assess contamination on a product before processing and then evaluate cleaning quality after the work is completed. The system became a practical decision-support layer inside the real operating process.

It combined image analysis, operator support, and quality control in one production workflow, helping teams make better decisions and improve consistency of results.

How the diagnostics flow worked

Before processing, the operator could take a photo of the product and let the system estimate the contamination profile. This created a practical recommendation layer for choosing how the item should be handled.

  • Photo-based assessment of the product before treatment.
  • Detection of contamination level from the image.
  • Identification of contamination character for better interpretation.
  • Operator guidance for choosing processing mode or parameters.

Quality control after processing

The second major part of the project was output quality control. After processing, the product could be checked again so the system could help determine whether cleaning quality was sufficient and whether the result met the expected standard.

  • Post-processing verification of the finished result.
  • Support for detecting insufficient cleaning quality.
  • Reduction of returns and rework through earlier quality detection.
  • More stable operational results across repeated production tasks.
Stack

Computer vision, messaging, and production workflow integration

The implementation combined image processing, operational data flow, and messaging between system components for use in a real production environment.

Osnovni stack

Python Computer Vision PostgreSQL Redis MQTT Eclipse Mosquitto Image Processing Production Integration

This stack supports image analysis, state handling, data storage, and message exchange between components in an industrial workflow.

Operativna vrednost

Operator Support Quality Control Cleaner Decisions Less Rework More Consistency Visual Inspection Process Guidance Applied AI

The result was a practical industrial AI module that helps personnel make better processing decisions and improves quality control as part of the operational flow.

Potrebna vam je pomoć? Imamo odgovore

Pregledajte najčešće postavljana pitanja i pronađite informacije koje vam trebaju.

Dobijate jasnu procenu trenutnog stanja, ciljnu arhitekturu i spremni plan implementacije. Svi deliverabli su dizajnirani za neposrednu upotrebu od strane vaših internih timova ili dobavljača.