- #Solar bird feeder how to
- #Solar bird feeder code
- #Solar bird feeder plus
- #Solar bird feeder download
Once I had the webserver setup I was able to download the photos from the SD card. It saved me a lot of headaches with trying to create my own, so thank you James!
#Solar bird feeder code
It was straightforward with only a few lines to integrate into my code and it worked great. I finally stumbled upon jameszah 's ESP32 SD File Manager. I spent 2 nights after work trying to figure out a way that I could do it with my meager coding skills. I figured there had to be a way to use Wifi to view the images in the webserver. I did NOT want to go out to the bird feeder and try to remove the SD card each time I wanted to see photos. The next area that I ran into trouble was how do I view the photos once they are stored to the SD card? Most tutorials that I saw just pulled the SD card from the ESP32 Cam and put it into a computer to see the photos. The Edge Impulse classify code is actually really straight-forward: //Perform Edge Impulse classification void classify() Once it is resized, the buffer is classified. That frame buffer is converted to RGB888 format, then resized to 48 x 48 pixels (the size that the Edge Impulse model is expecting). But the general process flow is that an image (aka a frame buffer) is received from the ESP-32 CAM. My code is kind of lengthy, so I won't put many snippets in the write-up, but you can see it on my github. I spent a lot of time digging through his code to see how it worked.
#Solar bird feeder how to
Louis Moreau from Edge Impulse has a pretty good tutorial on how to integrate an Edge Impulse model into ESP-32 Cam. You just add it as you would any library, and then include your inference header file in the Arduino sketch. Edge Impulse makes it easy to import the model in the form of an Arduino library. This is where I really appreciate the Arduino ecosystem. I then deployed the model to an Arduino library. Small enough that it could be run on the MCU. Then I mounted the solar panel to the roof of the bird feeder facing south (since I am in the Northern Hemisphere).įor pulling the model together pretty quickly, these are results I could live with. I mounted the camera using a 3D printed enclosure that I found (and had printed at my local library!) and double-sided tape to mount to one of the plastic housings of the bird feeder. Once I had the external antenna configured and tested on the MCU, I drilled a hole in the side of the bird feeder to feed the antenna cable through to the food storage area, where I was going to put the camera. I found that most boards have the on-board antenna as the default setting so you will have to do this if you intend to use an external antenna. But I got it done! I recommend you get a fine tip for your soldering iron and a magnifying glass to do this. The image above is magnified and those pads are tiny. With my rudimentary equipment (soldering iron with standard tip, no magnification) it was a PAIN to remove the resistor and put in the solder bridge between the two pads for the external antenna. Random Nerd Tutorials has a great tutorial on how to set up the external antenna that I followed. The ESP-32 Cam had a uFL connector already, so it's just plug and play, right? Well, not so much.turns out there are 2 settings for the antenna: the default on-board antenna (with 0 Ohm resistor) and the external antenna. Since I was going to put the bird feeder a good distance from my house, I figured I needed an external WiFi antenna for the MCU to boost the WiFi range. It's a pretty inexpensive microcontroller (~$10 with the USB adapter) but not a lot of memory (~520 kB) for image recognition models, However, it does have Wifi on it with the ESP32-S! I also had a ESP32 Cam lying around as well.
#Solar bird feeder plus
I had a 1W solar panel, plus the solar charger, the Powerboost, and a 2500 mAh 3.7V battery lying around. Once her masterpiece was complete, I took over.
You can see her finished product in the above photo. The project started with my 5 year old building, drawing on, and decorating the bird house. And with me being more and more comfortable with Edge Impulse, I figured I could add some image recognition as well! As usual, I ran into some hiccups, but overall I'm happy with how the project turned out. I figured I would give it a shot with a solar-powered version with a lower-powered MCU. I thought it would be a fun project to make a solar powered bird feeder camera so she could see the birds in her bird feeder! I've seen some image recognition bird feeder projects on here before, but they were with pretty hefty Raspberry Pi SBCs and hard-wired for power. For Christmas, my daughter received a "Build your own Bird Feeder" from her grandfather.
0 kommentar(er)
