So what was that word that you forgot but you now remember? Barista. They are the people who invent all the “weird words” on coffee menus, or so I heard elsewhere.
20160909/https://en.wikipedia.org/wiki/Barista
Questionable documentation. Docudrama. What is that anyways? Oh, I’m pretty sure it’s not about documentation, I was just fictionalizing that reference. It’s about documentaries, that’s something different.
20160909/https://en.wikipedia.org/wiki/Docudrama
20160909/https://en.wikipedia.org/wiki/Artistic_license
Here, take note. Professional microphone shock absorption mount.
20160906/https://en.wikipedia.org/wiki/Shock_mount
20160906/https://en.wikipedia.org/wiki/Microphonism
Ohhh… interesting.
20160906/https://munchies.vice.com/en/articles/youre-holding-your-coffee-mug-wrong-according-to-scientists
So what were you saying about processor interrupts? That it is not a problem that they can prevent useful progress. Actually, this is a recognized problem, you just had to search harder for the standard literature that names it. It is called “interrupt storm.” Although, one may be led to believe that the negative consequences of interrupt storms are entirely operating system problems, given practical rate limits in devices sending interrupts. Or is this not true? Is it really true that the hardware can generate interrupts in a way that is totally impossible for software to properly handle? Well, I don’t know, but the general problem associated with this is well-recognized.
20160903/https://en.wikipedia.org/wiki/Interrupt_storm
20160903/https://en.wikipedia.org/wiki/Advanced_Programmable_Interrupt_Controller
Okay, I’m wondering. What does the Internet have to say about insurance and accounting for libraries? Because, the thing here is, the library maintains its own inventory system, yet accountants and insurance agents each want to use their own incompatible inventory systems for their own purposes. So, what’s the remedy between the incompatibility between the disparate systems? Surely, there’s got to be a lot of information about this on the Internet! Unfortunately, I have failed to find very much.
Instead, what I found was information on this “new invention” of the modern era of “tool lending libraries.” Now, this is an interesting new addition to the chaos of the world, because in addition to the systems designed for book lending libraries, we explicitly have “new systems” “tailored to the needs of tool-lending libraries.” Yes, that’s where the “chaos” comes into play.
Maybe I should have searched for “Standard Operating Procedures for libraries” or something like that. But that is for another time. For now, I’ve got to catalog my findings in a more organized manner.
Oh, interesting. Reading about robots. Industrial robots versus personal robots. Unfortunately, much of the traditional robotics software is proprietary.
20160829/https://en.wikipedia.org/wiki/Personal_robot
20160829/https://en.wikipedia.org/wiki/Industrial_robots
20160829/https://en.wikipedia.org/wiki/File:Offline_teaching_welding_001.png
20160829/https://en.wikipedia.org/wiki/File:RoboLogix_Work_Envelope.jpg
20160829/https://en.wikipedia.org/wiki/Gimbal_Lock
This is somewhat interesting, but not as interesting as 3D scanning. Other things from the media computing group. Personal Orchestra, which is basically you are the conductor and you wave around an infrared wand to time-stretch a video of a performing orchestra.
20160830/http://hci.rwth-aachen.de/po
Okay, so I went back and read that Fabscan bachelor’s thesis paper, and I was impressed. I’m glad I did read it. It had a lot of useful general information on most modern related 3D scanning solutions, and it even included some new information on open-source 3D processing software that I didn’t know about. Specifically, the Powercrust 3D reconstruction algorithm provides for a much more elegant way to reconstruct meshes than the other more-laborious techniques I’ve been using in Meshlab. Also, the paper put a great deal of emphasis on the fact that the resolution of the 3D scanner is limited by the resolution of the stepper motors to advance the turntable and laser. This is important because up until this point, that was a complete oversight on my part on building high-resolution 3D scanners. I just thought that moving the laser by hand would be easy enough, but seeing what a problem it is even for motorized equipment, perhaps I should think again.
Does the Fabscan use any special techniques for higher resolution on the laser line generator? Actually, no, it doesn’t. Furthermore, it even goes to assume the laser line is a perfectly straight line, when in reality it isn’t.
One important point that the paper notes is that OpenCV has a camera calibration routine in it. Thus, there is no need for me to write my own camera calibration routine, unless I want to create a working solution with an absolute minimum code complexity.
- FabScan research papers.
Better design for 3D scanner. Okay, so you were noting that we have this problem where you want to be able to scan using a light source that doesn’t give you exactly a straight line. This is in fact a very good problem to solve, because in theory, even projection displays don’t give you a perfectly square picture either. So how do you do this? First of all, you need a sensor calibration stage. First you have to calibrate your camera’s angles (obviously), then you have to calibrate you line generator source. You do this using your camera and a flat sheet of paper. A first assumption that is very important. Your line generator or light source has to shine from a small object that can be considered a point light source. This makes subsequent calculations much easier. And in fact, when you think about digital projection displays, this same principle applies to then just like it applies to laser line generatiors.
So, you have a flat white surface that you know exactly where it is located and therefore know its exact 3D geometry, you shine your light source on it, which is situated at a known location and known angle, and you photograph the result. This allows you to turn that laser light source into a series of straight lines that emanate from your point light source. Then, during 3D scanning, provided that you know the position and angle of your light source, you can solve for more accurate point positions based off of the light intersections. But still, it is helpful if your light source is approximately a line/plane, for the purpose of simplifying the computations.
I tried setting up a Mattermost server and playing around with it. These are the commands I used to set it up, and some notes about other things I’ve found out about Mattermost.