From the microcosm to the macro, here's a couple headlines that are only related by their mention of solar panels.
Flea-sized solar panels embedded in clothes can charge a mobile phone
Dec 2018, phys.org
Team locates nearly all US solar panels in a billion images with machine learning
Dec 2018, phys.org
It sounds improbable that our clothes will one day power our electronic devices. But as our ability to draw electricity from the sun gets better, and as our devices demand less energy for more computation output, it seems inevitable.
The second headline reminds us that Big Data has found its match in Deep Learning. And this is one of the best examples, where satellites orbiting the Earth, their persistent gaze, from so omnipotent a vantage point, are generating data about us and our planet that we never thought we would see.
Beginning a few years ago we saw a similar thing perhaps even more ingenious - satellite images were used to measure the extent of infrastructure in regions without organized or reliable records for such things. A metal roof shines differently than no roof at all. And roads covered in asphalt (which contains tiny, sparkling glass pieces) will also shine differently. So the data is there. What I will call low resolution data, digested on such a large scale, becomes high resolution data.
Notes:
Infrastructure Quality Assessment in Africa using Satellite Imagery and Deep
Learning [pdf]
Stanford et al, 2018
No comments:
Post a Comment