Richard Körner - my email address - contact

New backend for my Website

I recently became more interested in web-dev. I in the process I found out that my old website ranked very poorly in the lighthouse report. See for yourself:

old lighthouse quality analysis report
The Report gave a lot of guidance on what to improve. The main switch I have done was that now I have full control over the server. Before I was forced to use the LAMP techstack but now I can use whatever I want. I used SvelteKit since I like the fact that it is generating the code to minimize the payload and the way javascript integrates with html and css seemed very familiar. It was kind of hard though as it expects you to be a professional to be able to use it. It was very messy at first and had to be reworked multiple times. The result was worth it though. I now also finally have https. This is the result:

newlighthouse quality analysis report


At OroraTech I was working on an orbit-simulation that was to be used for the development of wildfire detection algorithms. The goal was to create satellite imagery and testing how the present wildfire detection algorithms would perform on images from different sensor configurations.

This project has gotten it's last commit 2 years prior. A colleague has written a raytracer and set most of the systems in place. The reason we use a raytracer instead of a rasterizer was that we were able to mimic more physical effects.
Me and another colleague picked it up again. I was mainly responsible for the features (such as sun glint and texturing). In the .gif below one can see a 10° inclined orbit.

view from satellite as it flies over a simulated earth


In SLM-Processes the melting of the pores creates spatter. This material which should have been melted into the rest of the object is then missing and a pore may be left behind. Thus the reduction of such particles is desired.

The amount of particles depends on the process parameters such as scan speed and laser power. To determine which parameters produce less spatter one needs to quantify them. This is a difficult task due to many factors influencing the ability to track spatter. They can be of all kinds of brightness, background noise, speed can create trails and smoke plumes that glow so bright that it lights up the surroundings.

My work was to quantify these spatter using video material captured using a high speed camera that my colleagues collected for me. For that purpose I built a particle detector in python with opencv and a custom object tracker in rust. Here is a small sample:
highspeed capture of spatter from the 3d printing SLM process tracking of particles from 3d printing SLM process

Since the correct object detection is very difficult in itself, this process might be conducted by a trained algorithm in the future.