During the summer of 2006 I was employed by a company called Marel hf. Marel is an international leader in the design and manufacture of food processing equipment.
My project was to research hyperspectral imaging as improved means for finding imperfections in the color tone of food products. The project mainly consisted of setting up a hyperspectral imaging system and developing software to use it. Much effort went into developing computationally practical algorithms for distinguishing between different tones of color. We also wanted to learn about the behavior of different light sources with time especially in varying temperature environments so I conducted multiple experiments and analysis of those experiments.
The composition of the lens
The setup of the equipment, a line-lightsource, the camera and lens and a conveyor belt
Images captured with the system that show the spectral output of different light sources with time
A reconstructed RGB image from hyperspectral data. This image was reconstructed using an algorithm I devised and calibrated to get RGB data from hyperspectral data. Behind every point in the image is a spectrum of light, this is basically a 3D image seen from the top. Shows the significant segments from a salmon
