https://www.adafruit.com/product/2652

Despite the fact that it's 20 dollars for the sensor it might be worth updating the code to use a single bme280 in the center of the incubator box to check the environment - it can also sense barometric pressure if one is interested

I think it's worth the upfront price as only one i2c sensor is needed to monitor the whole incubator environment!

Look for models that could automatically do what we're now doing manually.

When assessing health label positions where discolorations occur.,

Implement the pipeline to take images from S3, run them through the pipeline, and store the metrics to a cloud database. The intermediate processed images will also be stored in S3.

Once done the scripts can be put on a server controlled by a chron job.

Looking at the code it seems that all the images are still stored on the RPi and also in s3. Should we delete them from the RPi?