iOS app for ‎autism speech therapy Multi-Camera Machine
Vision System

  • Client:
    R&D company (manufacturing industry), Singapore
  • Request:
    Develop a server for maintaining ZED cameras and transmitting scanned data to a database.
  • Result:
    Fully automated, fast, and accurate dimensioning; five times higher savings.
  • Technologies:
    • Qt C++
    • Boost C++
    • VTK C++
    • Point Cloud C++
    • HTML5

Challenge

The given project was organized as a collaboration between Cogniteq and the Singapore-based company engaged in computer science and scientific research to provide logistics and warehouse automation solutions.

The company got a great idea to develop a system for fully-automated, quick, and accurate measurement of cargo dimensions and weight. The project was aimed at the complete elimination of manual handling operations.

Our client planned to install a scanning system in the goods reception area of the production facility or the shipping area. The leading-edge technology makes sensitive dimensioning, weighing, and barcode reading with a simple press of the scan button on a handheld scanner.

Then the captured measurements are grouped and sent electronically in real-time to the web service of the warehouse management system database.

To bring this idea to life, the company needed to develop a server for transferring scanned measurements to a database. The Cogniteq team was invited to the project due to our vast experience in helping enterprises streamline diverse processes and optimize their expenses.

Solution

  • How the Machine Vision System Works.
    Upon receiving the information about the presence of the cargo from the METTLER TOLEDO weighing device, the system captures the dimensions of a cargo piece.

    Four ZED cameras (one on each side) make 3D cargo scanning and then send four color point clouds to the control computer.

    With a special algorithm, point clouds are combined into a single sensory image used to calculate the width, height, and depth of the cargo.
  • Data Management Server.
    The Cogniteq experts were tasked to build a server on OS Linux to maintain four ZED cameras and transfer the scan results to the DB Schenker connected with the warehouse.
  • Client-Side.
    It is a control program that runs on an Intermec wearable industrial computer. It assists operators in controlling data, operating modes, start, and endpoints of measurements.
  • System Features.
    Working with the machine vision system was only possible with a powerful graphics adapter that reads a stereo image from each ZED camera and converts it into data on the cargo dimensions. It is essential for mathematical transformations, in particular for building a point cloud, which is exploited for calculating the dimensions.
  • Setting up ZED Cameras.
    We developed unique software to make the configuration of ZED cameras as simple as possible. This is an extremely crucial functionality since incorrect camera configuration leads to inaccurate dimensioning.
  • Setting up Algorithms.
    The software has a fairly large number of settings. Users can choose any of the fifteen parameters to set up the measuring algorithm.

Result

  • image
    The machine vision system has been successfully implemented in the client’s warehouse.
  • image
    The company has achieved fully automated, fast, and accurate cargo measurement. The innovative solution has skyrocketed our client’s growth as now the measurement time and accuracy are 2-3 seconds and 0.5-5 cm, respectively.
  • image
    Thanks to the system, the measurement of cargo dimensions and weight costs the company five times less. Using ZED cameras as an alternative to laser scanning rangefinders was proved to be the best value-for-money outcome.