Industrial vision system
This training will enable learners to grasp the concept of the vision system chain, from acquisition to shape recognition in a context of industrial vision in order to be able to deal with many industrial applications, such as quality control, optical metrology, presence verification, etc.
The course lasts 5 days (30 hours) and includes:
- 18 hours of classroom-based learning
- 12 hours of practical computer-based work
Industrial vision has many applications in product quality control, a constant on the production line, as well as on finished products, their packaging or even their storage. It applies to a large number of industrial areas: automotive, mechanical, electronic, food and pharmaceutical, and is gaining many more sectors (aeronautics, textiles, seeds, etc.). In addition, recent developments in both hardware and software mean today we can benefit from applications that were not previously possible.
This course aims to present the different aspects of a vision system, from the acquisition of images to its interpretation through processing and analysis. The course focuses on the hardware aspects (descriptions, the technology and technique behind the image sensor, signal acquisition). It then outlines the main mathematical and algorithmic tools used for processing and analyzing images and presents numerous industrial examples.
Course level: Basic/Advanced
Basic mathematical knowledge
Some knowledge of signal processing
Basic programming knowledge of TP (Matlab/C language).
Frédéric LERASLE:
UPS Lecturer in Computer Vision and Image Processing,
Researcher at LAAS in Vision for Robotics.
- Introduction, overview of machine vision applications and the process chain
- Acquisition and sensors
- Image formation
- Optics and camera sensors
- Lighting techniques
- Improvements and image processing
- Radiometric, geometric and colorimetric corrections
- Principles of noise and linear and non-linear filtering of images
- Mathematical morphology tools
- Segmentation of images and descriptors
- Contour and region segmentation
- Descriptors of forms, appearances
- Classification and recognition of images and objects
- Prototype classification
- Parametric model classification: SVM, neural networks, Bayesian approaches
- Industrial vision applications
- Optical metrology
- Presence verification
- Compliance inspections
- Marker identification
- Overview of industrial vision software
- Specific cases during practical work sessions
Scheduled in French:
TOULOUSE: 17 to 21 May 2021
For the English realization, please, consult us.