Main components of the system:
1) Imaging system: Illumination units, camera, imaging optics and optical interface
Specifications:
Imaging unit coupling: The imaging shall be flexibly attached to any part of the mixing cylinder. An optional optical coupling unit can be used for optimizing image quality related to e.g. focus, which reduces the lens effect of the cylinder.
Illumination shall be designed to produce good contrast images of the bubbles. The illumination will consist of led and laser based illumination, depending on the material flow speed. Illumination and image capturing is synchronized with the rotator to consistently capture images with same location regards to rotator blades.
Bubble size distribution: Imaging system will be capable of imaging flow speed in mixing up to 20m/s. Image size and spatial resolution will be scalable from 2x2mm to 10x10cm with multiple options in between using interchangeable lenses. Possibility to enhance contrast in mixing with dyes and/or specific wavelengths will be available.
Bubble flow information: An alternative illumination unit will be designed for producing Particle Image Velocimetry –type of imaging that enables estimating bubble movement and to map bubble turbulence as a velocity vector field.
2) Analysis software
The imaging system is designed to fit the transparent cylindrical mixing unit of XAMK.
The scope of the imaging and image analysis is to provide information about parameters such as:
• Gas bubble size and spatial distributions as a function of time.
• Bubble flow characterization, e.g. bubble speed and direction per location.
Specifications:
AI/ML based image analysis tool: To provide the information related to bubbles, an image analysis tool shall be developed for this application specifically to enable automatic analysis of the image series captured by the imaging system, providing at least the following estimates.
- Bubble size and quantity information and derivates as a function of time: size distribution, D50/xx values etc.
- Bubble shape estimates.
- Designed to detect bubbles with different shapes, not only round ones.
- Bubble location information.
- Bubble flow and velocity information, similar to Particle Image Velocimetry
Analysis models to be trained/updated by end-user, using captured images.