Whew this is a rough one. There isn’t much to update. I had a setback with some code during the transition from AngularJS to Angular and decided to stay with AngularJS for now with the MEAN stack. I still have everything ready with the LattePanda which is producing the input and output for the project. I also had a lot of issues with OpenCV4NodeJS – not the library but doing everything I wanted it to do. I backed that off and have gone a slightly different direction. My OpenCV-fu is more Python-based. I love JS and would love to work this through Javascript. I don’t think, at this time, I’m going to. I’m gearing everything back towards Python for the OpenCV portion and then my renders to the site will be through a stream which Python will have already done the brunt of the work and then the feed will be the amalgamation of what I want it to do. There’s far more Python code out there to help with this.
I didn’t know about a tOLED (transparent OLED) that Sparkfun had out. So guess what’s on the way? Now, it is an OLED which means it won’t be able to display video like I really want. However, it will work for other aspects that I want it to. The AR/MR work will continue. The structure still remains at a display device (more than likely a larger phone for now) projecting data via HUD and transmitting/receiving data via Express. So the cameras will process data both in-phone and those external to a phone will be processed by the LattePanda and then using Express generate a screen representation for the phone to manage. All of the phone work is being done via React Native and VIROmedia. This leverages all of the functionality I really want with OpenCV in both the phone and the LattePanda. It also means that the phone can act as a display mostly and the rest of the work can be processed and displayed. That also means that some voice commands can be processed on the phone via its native STT (whatever that might be) and the LattePanda will handle the rest via whatever STT I choose in the future. I already have one set up via the web page display, but I might decide to change this later to a “native” version that runs on the LattePanda and forego the webpage version. I’m not sure, yet which direction that’ll be. I’d really like to get it all working and then make adjustments.
So, what will the tOLED actually do? The intent is to use the tOLED for daily work that can be done via that tOLED. That’s a roundabout way of saying HUD display text. Since there are so many aspects of this right now, the tOLED will be what I use for the in-between-HUD. I can use it in a smaller HUD that will display features I want in the interim – like speed, maybe displaying some text I want it to for questions I ask the AI. This could also include combining the sign recognition code I’ve already worked on to give a display that says what the signs say. I might also incorporate the STT for the AI to query what an object is or use it as a target selector, of sorts, to identify which items I’m looking at and which one I want to identify. There are a lot of possibilities just in the SERINDA framework that I can use the tOLED for. That was really long to say: I’m going to use it to display immediate items that need to be addressed where the AR/MR will be a fully functional display of everything. So this tOLED is the in between state. There are two tOLEDs that Sparkfun has available.
https://www.sparkfun.com/products/15173 – this is a much smaller version at $39.95 USD right now.
And the longer version https://www.sparkfun.com/products/15079 at $99.95
If you want a video where the tOLED is used by someone to make a smaller HUD here’s a video of that
I was also sent this last night for the Focals glasses. Which are about the same as the DIY version in display it looks like (meaning not rendering photo-realistic imaging). However, they are currently the cheapest version I’ve seen starting at $500 or so USD.
IN MY MIND BY WINK DOUBLEGUNS 