Earlier I wrote about our plans to utilize a neural network in the prediction of water quality throughout rivers and streams. I have done some more research since then so I will lay out some of our updated hopes and concerns. I have had some doubts about the amount of data we would need to build a model that can accurately predict water quality. Reading a few papers calmed my anxiety on that front slightly but I am still actively working on finding ways to incorporate larger data sets, which often provide a more accurate model.
As of right now I hope to leverage drone footage and iPhone footage coupled with water quality testing in order to build our model. Best case scenario would be a model that can accurately predict water quality through either drone or iPhone footage. Drone footage would allow organizations and park departments to deploy drones and capture imagery of areas that may not be easily accessible. iPhone footage through a future 4freshwater app, geotagged, would allow constant monitoring of water quality.
Updates to follow.
Some literature for those interested:
https://www.sciencedirect.com/science/article/pii/S1364815203001634#FIG1
https://www.sciencedirect.com/science/article/pii/S0034425702000093