Automated Assistance to Defect Detection

DiotaInspect

A set of tools for automated assistance to detection
and characterization of non-conformities.

Overview

For inspection tasks including numerous points to checks and with limited traceability, operations can be very time consuming, with potential high costs and delay due to late defect detection and poor visibility on the on-going process.

As a module in DiotaPlayer, DiotaInspect performs the analysis of the data captured on the field to establish a status on the elements to control. As a collection of algorithmic operators, DiotaInspect combines a set of automatic control features to detect and characterize non-conformities.

At the end of the check, a deviation report is made immediately accessible to the controllers, illustrating for each checked item its condition and the data used to rule on its compliance.

Main features

Various control techniques: geometric, colorimetric, recognition/reading, comparative analyses

Adapted to multi-configurations and various inspection cells: cobot, manual inspection with tablet, mobile workstation…

Detection of anomalies based on the digital model

Large control class catalog: brackets, harnesses, clamps, screw heads, cables, caps, etc.

Module integrated in DiotaPlayer or in the Digital-Based Robotic suite

Compatible with various types of cameras and sensors like the Sensor One

How it works

DiotaInspect workflow

Preparation of the job card: control classes, labels, tracking model

In-process field data collection: visualization of the step-by-step instructions to manually capture data in DiotaPlayer or automatic collection of data through digital-based robotic solution

Processing of data with Diota algorithms in DiotaInspect to assess status of elements to control

Generation and export of inspection report for further analysis

Applications

Different applications and benefits depending on your context, manual or robotic.

Human context, with Digital-Assisted Operator solution

Automated assistance to defect detection for DiotaPlayer in manual process:

For production, to rapidly detect possible non-conformities early in the assembly process and prevent late reworking, ensuring “first-time-right”
For inspection, to save time and allow alert escalation to detect potential non-conformities in case of multiplicity of control points and types

Robotic control context, with Digital-Based Robotics solution

Automated control of configuration during Digital-Based Robotic operations:

For automated inspection and documentation on complex products subject to variable configurations
For fully automated simulation and robot trajectory planning, process execution supervision, data acquisitions, analysis and arbitration

Control classes

For inspection processes, the solution includes a set of control features that automatically search for a wide diversity of defects using a variety of methods.

Geometric controls

Presence/absence of brackets
Positioning of brackets
Interferences
Inversions

Colorimetric controls

Detection of objects
Presence/absence of objects (washers, lids, caps, obturators…)
Markings
Paint areas

Recognition- reading

Texts (OCR)
Series numbers, signal plates datamatrix
Bar codes
Symbols, markings

Comparative analysis

Comparison by image recognition
Training through reference data sets
Machine learning
Deep learning
Image correlation