30 Oct 2019
Artificial Intelligence + Shrimp Farming – Osmobot Update
Written by Zach Stein

We are getting ready to go to the field with our camera-based sensors that use artificial intelligence to measure dissolved oxygen and other parameters continuously in fish and shrimp ponds. Here’s our latest update.

Problem with Existing Sensors

A quick background – we are building Osmobot because existing sensors for aquaculture ponds simply do not work. They are too expensive and are too difficult to keep clean, calibrated, and accurate in the field. This leaves millions of fish and shrimp ponds unmonitored today leaving them significantly less efficient and more variable than if they had a continuous monitor running in them. 

Osmobot – Only Sensor Built for Fish and Shrimp Ponds

We are building the only sensor system specifically made for fish and shrimp ponds. Osmobot uses a camera to take pictures of a series of optical sensor materials that change continuously in the water. We train the units to detect and see the differences in the materials using an artificial intelligence technique called Deep Learning where we simply show our algorithms thousands of pictures and their corresponding values and have it learn to recognize the corresponding changes in the chemistries. 

Our Latest Findings

We are headed to the field! Our models are trained and working great. We are excited to continue to feed them new data from the field so they can continue to improve. Pictures and videos will be coming soon.

Next Steps

We will be conducting local testing here in California and then heading to Ecuador. If you are a farmer there looking to upgrade your monitoring techniques, please reach out to us!