Just wanted to follow up and say this issue was solved with the tau v0.0.4 update. I no longer have dropped frames or bad data packets.
If you have these issues. I would update to the latest GitHub and update your computer make sure there are no updates waiting for your bios
@Lazar-Demin I updated the GitHub issues. Looks like my PC is the problem. What are the odds that the opencv modules used lean heavily on a graphics card ? My main PC is quite powerful but does not have a graphics card . That's all I can think of right now
@Zakrze were you able to run any of the other examples from the GitHub repo?
Do they stop at the same point?
Hey @Zakrze ! sounds frustrating. Are you able to run any of the other examples?
What browser are you typing the 127.0.0.1 URL into?
Do you have any spam protection plug-ins ? may need to disable for that URL
@Lazar-Demin added two github issues with links to videos on Google Drive.
Let me know what I can do to help, or any additional info you need. Feel free to email me directly if you want to work through this quicker. info@klsolutions.ca
@Lazar-Demin ok no problem
I thought it was my cable because the tai server also runs slow .
But now I'm not sure it's a cable thing
I have some really nice USB cables
I have tried nearly 20 different cables and I still get 98% rejected frames. What can I do?
Is anyone else having this issue?
Is there a cable that people are specifically buying for this ?
Is it possible to have the Tau send the same frame twice, albeit at a slower pace, and build one frame from the missing data in the two frames?
EDIT: The Examples all run well except for the recordRawFrames.py It throws frame errors.
"Data error, actual size: 38356, expected size: 38480"
@Justtraps
I downloaded from the python website . Not through the command line. If you are on Linux, some Linux distros don't have an up to date python package.
But you can always download it for Linux from the python.org website.
The link for it is in the starter guide link that Lazaar posted
Here is a simple program that I wrote to convert the frame 3d data into STL files. It is important to realize that this is finite point data. The stl is always going to be as sporadic as your data points. If you get really clear scans you will have really clear stl models.
This Version has no Filters built in, if you need to filter out near or far data you will need to code your own filter until I get around to it.
Feel free to use this however you would like and redistribute it as you would like. If it is being used commercially I would just like to know in what industry and with what goal.
The Read Me file should be clear enough for anyone in here.
https://drive.google.com/drive/folders/1VHtHBXHjGwjwbmcE0H1XzyLDbJJtROKx?usp=sharing
I will update this to a github link at a later date
I am working with the data from the camera to build my image into an STL file.
Everything seems to work except the data I am putting in is no good. Should read: Always debug before posting
This would still be nice information
a description of what points_3d is ?
my assumption
[X,Y,Z , R, G, B]
XYZ as triangulated locations based on the triangulated scan of the sensor.
RGB, values set in a range based on smallest and largest Z value
This frame looks fine as a RGB or Grayscale and I've made sure I'm working with a good frame.
@Lazar-Demin any thoughts?
I found that having too many versions of Python was a problem so I uninstalled all previous versions of Python and reinstalled 3.9.5 The latest and it was good to go
@ModusPraxis I haven't had a chance to start this yet. If you are interested in how I build my STL files I can show you in C#, Feel free to message me on discord sandbox#9273
@Lazar-Demin Thanks so much, looking forward to going through these.
Looks like I also need to turn on notifications from onion forum, wish I saw this sooner!
Hey there, I've been working on a project that uses point clouds and turns the resulting information into STL object files. It has been a lot of fun. I'm really looking forward to using the Tau camera to do more of this kind of work. Is there a batch of scan frames / point clouds somewhere I can use to test the modifications I need to make to my program ahead of getting the camera?
Thanks!
Hey! instead of playing with 120V how about we pull some power off of your doorbell transformer which is usually hanging off the side of your panel or mounted beside it. Usually these suckers do 12V/16V/24V AC. You're safe, and the setup you're making is safe. now whatever you use to get 5v will be cheaper and smaller.
@Tyler-Brown Hey Tyler,
I think using the time in seconds would be very useful for you.
On the arduino that controls both your lights and your sprinklers seperately. You can shoot the time from your omega to the arduino constantly.
You then have an arduino sketch that is waiting for those specific numbers to come in:
#for sprinklers
If seconds is >= (6am in seconds) and <= (6:05 in seconds)
Turn Sprinkler on
#for Lights
else if seconds >= (dusk in seconds) or <= (dawn in seconds)
Turn Lights On
@pvadam I had to change my country from US to CAN and that solved this whole thing
I feel like this should be PINNED somewhere @administrators