Varying Quality Images
Hi,
I am trying to set up a download script where I download a lot of images from several years ago based on a center latitude and longitude. Google's Static Maps API offers some high-resolution images, but they are not timestamped.
I wrote a simple function to take in a lat/long and output a GEOJSON coordinate space around that point. I then make a request to Planet with this geofilter and get a list of assets that I could use. I choose one, get its item_id, and then pass it into Planet's tile api (https://developers.planet.com/docs/api/tile-services/). However, the outcome is very inconsistent depending on the asset I chose. Some images are half-whitened out, and others are clearly lower quality than others. I've attached three examples of images of the same URL request but using different item_ids:
Observations
I used shapely to try to do some pre-processing, even before downloading the images. I realized that every time the image is cut off (like the first one), the geometry of the item_id (which is provided in the response of all item_ids that satisfy my filter) does not even contain the original shape I created.
I have not been able to come up with a way to definitively identify the middle image over the last image pre-download. This would be useful because the middle image is clearly a better option.
Does anyone know why this is the case? Shouldn't the returned item_ids be guaranteed to at least contain the space you give? And how can I make sure to download the middle image and not the last one without downloading both and choosing the better one?
Please sign in to leave a comment.
Comments
0 comments