Elf on the shelf

Elf on the shelf

Ok, I’m going to trying to do a dictation of my new blog entry using this software called VOSK. V O S K. Anyway, it’s been eighteen months since I’ve done an entry and a lot has been going on.

I’ve done a few things with IoT devices. I’ve been replacing the factory shipped firmware on these IoT devices with Espurna and Tasmota. This has involved opening up factory sealed devices and making physical changes to allow reprogramming. I have also leveraged OTA firmware updates. I was pleased to discover that the Arduino IDE had added the OTA feature, all though I have yet to try this route. I typically have been loading firmware with Python, using ESPTool.py. I found a great video demonstrating how vulnerable these devices are, take a look here, https://www.youtube.com/watch?v=g3o_-AiswF0. It’s also worth mentioning much of my work was based of the information provided by these folks. https://www.youtube.com/c/DrZzs/about and https://www.youtube.com/c/digiblurDIY/about. This led me to look more into the MQTT protocol and services like Node-Red and Mosquito.

Earlier this summer I have done some more video post processing. Video footage from drones and radio controlled planes with cheap key fob cameras pushed me in this direction. Those key fob cameras were my go to when I was doing a lot of bicycle riding, something I hope to do more of soon. Post processing was done using FFMpeg and ImageMagic. The aerial footage was mainly interpolated, stabilized, with some edge sharpening. I also did some underwater footage where normalization and histogram balanceing cleared up the view. Our summer was mired with forest fires, so I also was able to apply the same with aerial footage.

I’ve done a little audio processing, mainly text to speech and this example of speech to text. The tts has been primarily espeech, which may sound like the late Stephen Hawking’s SGD to the casual listener. I’ve been using it to announce or dictate feeds using calls. As I mentioned at the start of this, vosk has been the stt, which is a Python based API.

Also, since I was doing aerial videos with remote controlled planes, this has moved me more into this direction. Especially with the loss of one of the birds. That event was not costly fortunately, but the lemons were good fruit for a sweet drink. Loosing the plane led me to research location or position technology. BLE iBeacons were the main focus of my efforts. I had purchased ESP32 Cam modules not long after my last post in 2019. These devices fit perfectly into that project with on board support for BLE and data storage built in. They can also take serial feeds from a GPS device so a heat map can be generated to locate the device. I’m still grounded.

I’m also using software defined radio and python to detect third party 433 Mhz devices, such as motion detection, door sensors, and temperature and humidity sensors. This has worked back into my work with MQTT as well as Espurna firmware on a Sonoff 433Mhz to Wifi bridge. I think I’ve found serious limitations on the reliably of setup, which prompted me to abandon using it to control cooling fans in my attic. That job is still being done reliably to a simple thermostat, proven technology. At any rate the sensors have been easy to install and maintain. Now I have more visibility with the Node-Red web UI these sensors feed into.

Some other IoT controls have caused me to dust off my ESP01 modules to control LED color strips. I’m likely to use these more with my lingering embedded projects, like my RC plane or bicycle helmet. It’s my intent to do projects where the work can be reused. I’m finding it more of a challenge than ever to find free time. This is the main reason I haven’t posted much is so long. It was either write about ideas or work with something real and write about it another time.

At least now I have a post for this year, looking forward to the next.

Comments are closed.