Algorithms and ideas on measuring 3D particles with cost efficient hardware

Measuring the volume of dirt particles within the context of technical cleanliness is an important task. The estimated 3D shape yields significant information about material properties and can highlight potential risks while using the produced goods within the final product. The results of this analysis are used to optimize the production steps to eliminate such harmfully shaped particles already in the early stages. Estimating 3D volume of particles with just a few microns of surface area normally implies an expensive microscope measuring device. This report investigates on hardware and software approaches to implement a cost efficient sensor consisting of standard components. A hardware setup using eight LED lights for illuminating the specimen from various known directions while using the shadow casts to reconstruct the object surface are proposed and evaluated. By manipulating the intensity of these LEDs the setup can be adjusted to be useful for various specimen with different reflection properties. The height of a particle is estimated at discrete positions based on a global model. Measurement results acquired by established state-of-the-art microscopes are analyzed for incorporation into the donated model for initial ground-truth. During the calibration process those results are used to map measured height values to the scalar z-values used for height estimation of the given specimen