TIA QualityHealth consists of applications where images are evaluated using deep learning algorithms that are focused on defect localization and classifications, near real-time image analysis is performed on streaming data and defect detection algorithms are used.
The first step for the defect localization is the use of related dataset. At this stage, TEKNOPAR image processing library is used. Images of defective and defect free products are collected by industrial high-resolution cameras. The library contains functions for data augmentation, resizing, imperfect data identification, image enhancement, ROI extraction, evaluation, and labeling appropriate for their use. According to the requirements of defect detection, deep learning, classical image processing or statistical methods are developed. Defects are localized and evaluated by the algorithms in the library, and the results are presented to the users via web-based applications. Being a flexible system, TIA QualityHealth can be customized according to the specific needs and requirements of the users.
IPR / License
Commercial. All rights reserved by TEKNOPAR.