Select to view content in your preferred language

Troubleshooting: High Resolution Landcover Classification, Pre-trained Deep Learning Model

200
1
11-20-2024 07:45 AM
Labels (3)
InukSimard
New Contributor

Hi all,

 

I'm getting this result (Figure 1) from running HighResolutionLandCoverClassification_USA.dlpk using "Classify Pixels Using Deep Learning" on Figure 2. Any insight on why this is happening would be appreciated. Figure 3 shows arguments, Figure 4. Environments.

1. Output Raster Dataset (Value 0 to 254):

InukSimard_0-1732117058658.png

Figure 2. Original image (RGB, 0.2m resolution, 8-bit unsigned)

InukSimard_1-1732117203495.png

Figure 3. Parameters: Arguments

InukSimard_2-1732117430493.png

 

Figure 4: Parameters: Environments

InukSimard_3-1732117485238.png

Thanks!

 

Inuk

 

0 Kudos
1 Reply
PavanYadav
Esri Regular Contributor

@InukSimard the output appears to be empty (likely nodata pixels). Can you try a small area with:

  1. GPU (don't use Parallel Processing Factor, it's for CPU)
  2. CPU without parallel processing factor 
  3. CPU with parallel processing factor 

Please share which of three works for you. Also, please share what version of ArcGIS Pro you're using. 

Pavan Yadav
Product Engineer at Esri
AI for Imagery
Connect with me on LinkedIn!
Contact Esri Support Services
0 Kudos