Technical Blog

Along time between posts

So it’s been quite awhile since I posted anything to any blog. I did post a couple updates as I did projects in grad school. The Wordpress site I used for this earlier died and was hosted on a VMware ESXi server in my homelab. I didn’t have time to fix it while in school so no postings since mid-2016.

I’ve since graduated with my Masters in Interactive Intelligence (Machine Learning) from Georgia Tech in 2019 and picked up a couple of jobs that consumed the bulk of my time. So resurrecting and running my hobby website just got shelved. I did some work in NLP and wrote an academic paper with a PhD Candidate who became a PhD in Linguistics. I have a couple YouTube videos, a conference presentation and some code from that work which I should write an article to share some of the experience and code.

Part of the problem was trying to figure out the best way to get a relatively static website up and running that would not take too much effort to maintain and manage. Analysis paralysis took over for awhile as I looked into serverless options and various hosting providers.

At one point, I considered building a WordPress site again on a Digital Ocean droplet but I was facinated with the whole static website generated from YAML or MarkDown files but it seemed limited or challenging to maintain. After recently taking a quick Jekyll online lesson on Github and doing some digging around, this seemed like a path forward that offers a lot of options. This requires learning from gray areas of development I’ve avoided but I’m coming up to speed on this stuff which is a good thing.

The Github Actions component of this website build is facinating and I’m still wrapping my brain around what I can do with it. I did some side project work on OpenLens custom builds for awhile to support a work initative and it was very powerful.

So I’ve dug out my old WordPress website archives, some articles from my Google Blogging and even older archives for DarkMagic.org that includes some work from 1997 when I was at NC State University. That’s the majority of what I’ve posted here. I’ve got some additional content from the late 1990’s and early 2000’s that I want to clean up a bit and post still. Those I’ll dig into as I get time to do them. They are mostly C programming or low-level hardware work I did in those days.

I still have some rough edges with pagination, color schemes, the integration between the main site and resume site, but overall this came together quite well I think.

I’m going to try to make myself post something at least twice a month. I have two articles in DRAFTS that needs some polish before releasing and several more that I want to write on various topics. Maybe I’ll write a post on what I want to write about at some point later once I get a couple other posts out the door.

Cheers from the Carolinas and hope you have fun reading some of these brain dumps.

–Michael

PhotoSynth Export and Visualizer

I have developed something interesting for my graduate class in Computations Photography for the final project. As background, for the class we had an assignment in which we used the Microsoft Photosynth service to generate a 3D walkthru of an area by uploading several hundred photographs. On the back-end the Photosynth web service does feature extraction on all the photos and then related the photos in three dimensions to each other and the feature points. This generated a point cloud of related points between the photos.

PhotoSynth

Artificial Intelligence for Robotics (CS8803-001)

Artificial Intelligence for Robotics (CS8803-001)

Associated with Georgia Institute of Technology

Fall Semester 2015

The goal for the final project in CS6475 AI for Robotics was to create a robotic platform to investigate computer vision technology. The platform included an Arduino with sensors and motors and a Raspberry Pi 2 for the vision and primary control system. The project URL is a video channel that shows the progress and challenges.

Raspberry Pi 2 built-in LED

For an assignment in my robotics class, I need to have an autonomous system react to the environment around it. Reacting can be as simple as flashing a LED if a sensor detects a change.

I have two objectives for the Raspberry Pi 2 (RasPi2) and those are to take a picture using the 5mp webcam and flash a LED. I could use the standard GPIO pins and setup a separate LED but noticed we have two perfectly good LEDs built into the board.

Reading on these built-in LED did not elicit any clear way of interacting with them from the regular Linux documentation. I informally called them the Red Power and Green DiskIO LEDs. It was by reading the headers to the source for Windows 10 for Raspberry Pi 2 that I found the GPIO pinouts for these two LEDs. They are:

35 Red Power LED
47 Yellow DiskIO LED

Posts