Nepal: Planet's pre-event composite mosaic

[cgiovando]

Planet Labs has generously provided access to the their Nepal archive to use in OSM. They will also provide fresh imagery as soon as clouds clear up.

I started looking at samples from their archive and trying to understand what we can extract in terms of OSM mapping. Please add your comments below. I also would like to understand the quality of co-registration, so that we can advise in case of shifts with current data.

A composite and radiometric balanced mosaic has been requested, but in the meantime we may have to work with individual scenes. Here are some screenshots:

Having a pre-event mosaic would also allow us to run automatic classification and change detection algorithms to extract e.g. landslide areas. The output could be used in a dedicated TM task to guide crowd-sourced validation and more detailed mapping.

[cgiovando]

For reference, here's their archive coverage in the Nepal area

[pierzen]

First look,very interesting. What is the image resolution?

[cgiovando]

First guess I would say it varies between 3 and 4 meters. Here's some post event samples:

https://github.com/hotosm/imagery-requests/tree/master/Nepal/Imagery/Planet%20Labs/201504_POST

[cgiovando]

Here's the TMS endpoint for Planet's Nepal mosaic, mostly pre-event scenes:

https://tiles{0-3}.planet.com/v0/mosaics//{z}/{x}/{y}.png?api_key=

I will send API key to the activation coordinators to add the TMS to a new task. Signed license agreement is also here, let me know if you see any problem with it. They also clearly stated through emails that we can freely use it in HOT/OSM for Nepal EQ response (a-la NextView license).

Screenshot below, thanks Planet!!