You are required to read and agree to the below before accessing a full-text version of an article in the IDE article repository.

The full-text document you are about to access is subject to national and international copyright laws. In most cases (but not necessarily all) the consequence is that personal use is allowed given that the copyright owner is duly acknowledged and respected. All other use (typically) require an explicit permission (often in writing) by the copyright owner.

For the reports in this repository we specifically note that

  • the use of articles under IEEE copyright is governed by the IEEE copyright policy (available at http://www.ieee.org/web/publications/rights/copyrightpolicy.html)
  • the use of articles under ACM copyright is governed by the ACM copyright policy (available at http://www.acm.org/pubs/copyright_policy/)
  • technical reports and other articles issued by M‰lardalen University is free for personal use. For other use, the explicit consent of the authors is required
  • in other cases, please contact the copyright owner for detailed information

By accepting I agree to acknowledge and respect the rights of the copyright owner of the document I am about to access.

If you are in doubt, feel free to contact webmaster@ide.mdh.se

A Mobile Machine Vision App with VLM-based Validation for Industrial Inspection Data Acquisition

Authors:


Publication Type:

Conference/Workshop Paper

Venue:

2026 19th International Conference on Machine Vision.


Abstract

Reliable 2D image-based dimensional measurement depends not only on the measurement algorithm but also on how the image is captured. In field inspection, mobile phones are practical and flexible, but uncontrolled handheld capture introduces changes in zoom, distance, viewpoint, framing, and motion blur. These variations make later pixel-to-millimeter conversion unreliable. To address this problem, we propose MetricCapture, a guided mobile acquisition app for preparing measurement-ready inspection images. The app is developed for an industrial excavator sprocket-inspection case study. The proposed workflow combines capture-control constraints with a two-stage AI validation strategy: a YOLO26 detector performs fast pre-capture object-presence validation on the live camera, while an on-device Vision-Language Model, SmolVLM-Instruct, performs prompt-based post-capture semantic validation to verify that the captured image contains the required sprocket structures and visual context. The app fixes key camera settings, guides the user using angle and framing feedback, rejects invalid captures, and stores accepted images with structured metadata for downstream dimensional measurement analysis. The YOLO26 detector achieved 0.99 precision and 1.00 recall. SmolVLM achieved 94.0% post-capture validation accuracy, with 6.0% false acceptance and 6.0% false rejection, while the average post-capture processing time was 3.2s. The results show that combining guided acquisition with YOLO26 and VLM validation can reduce invalid data at the source and improve the consistency of mobile images prepared for later industrial metrology.

Bibtex

@inproceedings{Rehman 7465,
author = {Anwar Ur Rehman and Shahina Begum and Mobyen Uddin Ahmed and Shaibal Barua and Daniel Aurel and Emmanuel Weiten},
title = {A Mobile Machine Vision App with VLM-based Validation for Industrial Inspection Data Acquisition},
month = {December},
year = {2026},
booktitle = {2026 19th International Conference on Machine Vision. },
url = {http://www.es.mdu.se/publications/7465-}
}