SIMULATION
Designing inspection systems in a virtual world
Petra Gospodnetic, at Fraunhofer ITWM, describes her work building a virtual image
processing environment to simulate the design of an inspection system
P
roducts come in all shapes and sizes, requiring inspection system integrators to adapt or completely change their
machines with each new application. Tere is no one-size-fits-all vision system; specialised production lines require specialised inspection. It is a complex development process, which works, but it doesn’t come cheap. However, what happens when a client
requires an inspection system for a production line manufacturing a number of small batches of products? Tere is no smart solution available and if Industry 4.0 dictates an increase in production flexibility, together with a reduction in overall cost, how well can automated inspection keep up? Can costs be cut, inspection systems made less rigid, and the quality of the product improved, while making the integrator’s job easier?
What is preventing automated inspection? Development of a new inspection system is an iterative process. Te pre-study phase is used to develop and adjust the system until it meets specific requirements. When those are met, the prototype goes into production to clean it up and make it ready to work 24/7. It is the pre-study stage that could be
looked at closer. Te system is developed in two phases: image acquisition and image processing, with most of the development resources put into image processing. Hardware components and their set-up are decided by an engineer based on physical testing and a
A list of viewpoint candidates (white), along with the reduction of viewpoints required to cover an object’s interesting regions (blue)
trade-off between features and cost. It takes a lot of time and effort to test different hardware solutions, and it is impossible to test every potential scenario. Terefore, the engineer chooses what they know will work, even if it has certain drawbacks, and doesn’t spend too much time experimenting, because the measurement unit for hardware setup takes hours to change. Soſtware engineers working on image
processing are expected to make their algorithms capable of compensating for potential image acquisition weaknesses. For surface inspection, computer vision research is mostly focused on robust pattern classification, overlooking the need to optimise the acquisition design in order to distinguish those same patterns better. Today, robust classification of difficult patterns can only work in a highly controlled and rigid environment, where as many variables as possible are fixed. To enhance vision systems, these rigid image acquisition constraints must be loosened.
Closing the research gap Using computer vision, computer graphics, machine learning and robotics it is possible to build a framework capable of design optimisation, which removes the need to
60 Imaging and Machine Vision Europe • October/November 2018
A virtual image
processing framework … makes optimisation of component positioning possible, without actually requiring the engineer to remount the equipment
assume a fixed image acquisition setup. Currently, very little or no research is focussed on inspection system design and optimisation. A virtual image processing framework can
overcome this gap in research, by thoroughly testing the acquisition hardware of choice and simulating the end result. Most importantly, it makes optimisation of component positioning possible, without actually requiring the engineer to remount the equipment over and over again. Furthermore, computer vision algorithms can be developed and tested on simulated images, along with the acquired ones, overcoming a frequent problem of defect sample acquisition, especially in industries where defects occur rarely, but are critical when they do – aeroplane blisks and car brakes are two examples.
@imveurope
www.imveurope.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72
