I'm trying to do a supervised classification in ArcGIS Pro that will allow me to find Tar and Gravel rooftops in Miami.
I have Landsat data but just wondering if there is imagery out there with better spatial resolution that might work better?
-
Plant has 3m and 72cm daily 4-band imagery: planet.com/products/monitoring– Aaron ♦Sep 25, 2019 at 17:34
Add a comment
|
2 Answers
Sentinel-2 has a higher spatial resolution (10x10 m in VIS+NIR) compared with Landsat series (30x30 m, 15 m in panchromatic), so your results must be more accurate. Also, is a free to use sensor.
Here you can compare spectral resolutions:
answered Aug 1, 2018 at 16:28
-
I'm a bit new to this. I know spectral resolution is pretty important for image classification. Does the Sentinel 2 have as good a spectral resolution? Having a bit of trouble reading the image you posted. I know Landsat 8 has 8 bands. From what I've read Sentinel only has about half that.– BeauAug 1, 2018 at 16:35
-
@Beau Sentinel-2 has some bands in 10m, 20m and 60m. In 10m resolution, has RGB + NIR; in 20 m, has far red (a.k.a. red edge) + NIR + SWIR; and in 60 m, has some bands to retrieve quality data. Aug 1, 2018 at 16:37
-
Do you think RGB+NIR will be good enough for ArcPro to tell the difference between Tar and Gravel rooftops and everything else or that I may need more bands?– BeauAug 1, 2018 at 16:41
-
Spectral profiles of this kind of material are almost a line, so you should be ok. But 10 m isn't enough to compute differences on small buildings but is the better spatial resoution found on free imagery Aug 1, 2018 at 16:49
-
Thanks for all the help! All the rooftops I'm looking at will be industrial and so should be a bit larger. Also they tend to be all one material. I'm hoping to get at least a sample of what should be Tar and Gravel, then do some ground truthing with Google Earth.– BeauAug 1, 2018 at 16:52
Within the United States there is the National Agriculture Imagery Program (NAIP), which has 4-band imagery available every few years at 1m resolution. If you're using an ESRI product, it is very simple to load/analyze/export this data directly from the United States Department of Agriculture (USDA) Aerial Photography Field Office (APFO) via their Image Server: https://gis.apfo.usda.gov/arcgis/services
Note: This isn't a link that will open in a browser. You must add a new GIS Server in ArcGIS and point it to this address.
I have had great success with supervised classification between rocks and asphalt using NAIP and other 4-band imagery in the past.
