Search
Realizations

Woodspect for quality control of lamellas – FURNITURE PRODUCER

The project implemented for one of the factories of the multinational conglomerate designing and selling ready-to-assemble furniture, kitchen appliances and more.

Initially the order included 1 system installation in 1 factory. The solution was suppose to optimize the process of the pine lamellas’ repairs.

The main difficulty in automating the board repair process is the detection and classification of defects (e.g. distinguishing healthy from decayed knots or detecting resin pockets).

To solve the client’s problem, at KSM Vision, we recommended installation of the Woodspect system on the client’s production line. Woodspect is equipped with an optical scanner for recording color photos and the shape of lamellas and logic using computational algorithms based on neural networks (deep learning AI). Thus enabling the classification of the wood defects and detecting the wood lamellas requiring repairs.

Woodspect hardware and software (cameras with illuminators, marking systems, GUI, logic, etc.)  proved reliable and precise. In result the client decided to implement 2 more Woodspect installations in 2 other factories.

  • Solution Woodspect
  • Client Furniture Producer
  • In our opinion, the use of neural networks and the function of classifying detected defects distinguish the KSM Vision solution from competitive solutions, and the effectiveness achieved by the system ensures quality control at the level of systems of leading manufacturers.
    MLEKOVITA, 1 of 100 biggest Polish private companies 2022 according to Forbes
    M.A. Engineer Dariusz Sapiński,
    President of the Board
  • When designing the solution, KSM Vision showed great flexibility in adapting the system to the expectations of the BIOFARM Production Department. (...) Since the implementation of BLISPECT vision technology on the 1st BIOFARM production line almost two years ago, the KSM Vision Service Department has been available to BIOFARM employees with a good reaction time.
    BIOFARM Jarosław Pieczuro,
    President of the Board
  • KSM Vision Sp. z o. o. has developed a system for quality control of glass vials for pills. (...) A big advantage of the system is the use of machine learning algorithms, which allows easy adjustment of the system.
    ADAMED Group Dariusz Stępień,
    Dyrektor ds. Infrastruktury i Mediów

Zakres realizacji

The Woodspect system integration service included:

– development of the client’s own neural network architecture, which ensured appropriate performance and classification level and enabled the system sensitivity to be controlled using ‘tolerance’ parameters for each class
– the process of the system AI training based on the wood samples delivered by the client (identification and distinction between correct marks of the natural wood boards and the defects)
– installation of a Woodspect system on the production line
– guarantee and tests
– maintenance & support

Our R&D on the solutions with the use of neural network potential are backed by our renowned partners: