Systems with automatic defect recognition
R&D team of ATG Group develops own ADR (Automatic Defect Recognition) systems for automatic ilumination, capture, and indication (defect) detection and classification for methods VT, MT and PT. These ADR systems are delivered on customer ́s request also with manipulation including bin picking.
iVT systems allow detection of what a human operator may see and recognize, which usually includes wrong signatures of parts, scratches, pressure marks, corrosion, porosity, non-wrought parts or burnouts, cracks, wrong dimensions, wrong assembly, presence of foreign objects etc..
iMT and iPT systems allow assessment of presence, type, position and size of indications, which is a challenge itself due to complexity of shape, background level and other. ATG Group resolved this problem thanks to long-time experience in NDT of its staff that was reflected during implementation with modern and unique hardware and software (deep learning).
Registration level of main indication or defect dimensions typically reaching 0.1mm, but is further conditioned by the surface roughness.
Evaluation speed is always same or better than by human operators (especially on larger areas the ADR systems provide significant advantage).
Systems are therefore able to replace the human operators for testing of parts´ quality and may reach more productive results than human operators.
In all cases the cooperation with customers is a key, especially important is provision of sufficient number of representative samples that allows precise indication or defect recognition via deep learning process.
Complexity
Complexity increases with complexity of parts and complexity of its surface and roughness. For that reason the suitable candidates for automation with ADR are mass produced parts with higher surface quality, where the expenses associated with deep learning process and cost of equipment are lower and thus the return on investment may be reached sooner.
Return on investment is usually in the range from 1 to 3 years depending on the testing actions to be performed on the given part, labor cost in the area of installation and also current productivity of existing processes.
