Hello!
I am happy to be part of this challenge, and also very happy that my wife will be adding her artistic inputs to make this an even more fun project 
The project kit we were sent for this challenge includes an Arduino MKR WAN 1310 and a Nicla Vision. The MKR WAN board should allow us to connect using LoraWan, if a receiver is within range, and the Nicla Vision board, with its camera and onboard support for TinyML should allow us to detect objects and things, and report on them.
For this project we intend to use computer vision to detect bees on fruit tree blossoms, as well as detecting invasive murder hornets, aka Northern Giant Hornets.
These giant hornets are known as murder hornets because they prey on North American honeybees, which have no defenses against them.
Murder hornets have already been found just over 100km South of where I live (on Vancouver Island, BC, Canada), and it would be beneficial to have automated stations set up to help with early detection, so that authorities and volunteers can intervene more quickly when necessary.
We also want to detect bees on blossoms to help determine whether the timing of the blossoms aligns with the timing of the pollinators.
One problem in the last few years has been with the weather being warmer or colder than expected and bees either not being available when trees were blossoming, or the other way around with bees not having enough blossoms when they need them.
These are our project goals:
- Detect murder hornets.
- Count the number of bees
- Make the results available via a web browser
Stretch goals include:
- Identify various bee and wasp species and estimate their relative populations.
- Detect non-bee critters like moths and flies.
For this challenge we will use the hardware kit provided to detect artist created “insects” on a fake blossomy tree we have.
As this is a winter project, we need to use artistic stand-ins for the prototype, but the only change that should be needed to make it useable in the wild would be re-training of the computer vision engine.
