Products

MathTechStudio

Our software framework for optical quality control integrates with the HALCON machine vision library by MVTec and with the open source library OpenCV. MathTechStudio makes it possible to develop and deploy individual, automated inspection solutions.

Request a demo

Inspection solutions built around your process

MathTechStudio makes it possible to develop and build customer-specific, automated inspection solutions.

Anfrage senden

Online inspection in live production: wave profile, defect list and the assessment of the coil by quality grade.

Three modules, one clear structure

Online inspection module

Runs as a DLL alongside the customer's machine software, or as a complete application on the customer's system. This module has access to the database in order to exchange data, inspection results, images and AI models.

Offline labeling tool

Used at the customer site on a separate PC with access to the database. Inspection results can be reloaded, assessed and labelled here. Its main purpose is to state and communicate what production actually requires, so that the development of a new AI model can be aimed at it.

Offline analysis tool

Used to train the AI models and to configure the online inspection module. Besides training and optimisation, the individual components of the AI can be validated in an offline test environment, to confirm their quality and to examine in detail how a particular decision came about. Training can be run at the customer site on a high-performance PC or at MathTech.

The workflow

Inspect

Online inspection modules analyse and inspect the images using HALCON and the available AI models. The results are then passed to the machine and stored in the database.

Label

Using the offline labeling tool, the customer reviews the results and marks the relevant images or regions as a fault or not a fault, in various classes. That information is stored in the database as well.

Train

All information, images and labelled data from the customer's experts or from the MathTech engineers are loaded into the offline analysis tool, and the AI models are trained from scratch or trained further.

Optimise

The optimised models are installed back onto the online inspection modules and used for live inspection.

Flexible, extensible, scalable

The three modules of MathTechStudio together form a flexible, extensible and scalable solution across several production sites.

The MTS framework is a single environment in which hardware, datasets, algorithms and software come together into one complex machine vision solution.

How an inspection works

Features are derived from the images and then weighted and classified automatically. The analysis tool makes both traceable: which features were found, how they are rated, and which class follows from that.

The MathTechStudio analysis tool: inspection jobs, defect classes, AI configuration and a surface image with detected defects marked in green — The analysis tool: assessing inspection results, assigning defect classes and configuring AI models.

Video analysis with HALCON algorithms

MathTechStudio has been extended. Video data, from files or from streams, can now be loaded directly. It is also possible to process video data straight away with HALCON algorithms, in real time or offline, integrated either directly or as HALCON procedures through our machine learning interface. All current HALCON versions, such as the HALCON Progress and HALCON Steady releases, are supported.

Monitoring of production and of warehouses
Detection of moving objects such as cars, bicycles, aircraft and ships
Processing of drone footage

Free and without obligation

Send us your inspection task. We will tell you whether it can be solved.

A specification, a defect catalogue, samples or images are enough. In a first free feasibility assessment we evaluate the task, settle the open questions and plan the next steps with you.

Have the feasibility assessed sales@mathtech.de

Consulting, training and support are part of the deal.