AWS NAIP open data

[kylebarron]

May be useful to check out AWS's NAIP imagery. https://registry.opendata.aws/naip/

You'd want the naip-visualization dataset (arn:aws:s3:::naip-visualization)

The data in this bucket is 3 band RGB in Geotiff format. It is converted and managed by Esri. This imagery is provided as Cloud Optimized GeoTIFFs, has been compressed using YCbCr JPEG with quality 85. It is provided as 512x512 tiles, with pyramids created using 2x sampling by averaging. The bucket is set as a Requester Pays bucket.

[kylebarron]

AWS NAIP is actually pretty cool because they have multiple years and multiple sets of data. The data in the naip-visualization bucket looks best for my use case:

The data in this bucket is 3 band RGB in Geotiff format. It is converted and managed by Esri. This imagery is provided as Cloud Optimized GeoTIFFs, has been compressed using YCbCr JPEG with quality 85. It is provided as 512x512 tiles, with pyramids created using 2x sampling by averaging. The bucket is set as a Requester Pays bucket.

NAIP annual coverage map

Directory Structure

• Imagery : Bucketname/state/YYYY/resolution*/bands/quad/.mrf (.idx, .lrc) files
• Metadata : Bucketname/state/YYYY/resolution/fgdc/quad/.txt files
• State/Year Index : Bucketname/state/YYYY/resolution/index/.shp (.dbf, .shx, .prj, .sbn) files

Although I've found that with naip-visualization, the imagery bucket is named rgb/ not bands/.

Process

So the process would be:

1. Take geometry
2. Load index shapefile, at e.g.

   s3://naip-visualization/wa/2013/100cm/index/
   

3. Intersect geometry with index
4. Generate map filenames. It looks like the USGSID, FileName, and ST columns are necessary. They have values of, e.g. 4812247, SW, and WA. And correspond to file names like:

```
s3://naip-visualization/{ST}/2015/100cm/rgb/{USGSID[:5]}/{FileName}
s3://naip-visualization/wa/2015/100cm/rgb/45122/m_4512229_ne_10_1_20150728.tif
```

5. Download all imagery for that year and state and tile it?

[kylebarron]

This is superseded by https://github.com/kylebarron/serverless-aerial-imagery