In ArcGIS Pro, for extracting building footprints from raster using detect object deep learning tool with extracting building footprint modal(DLPK). tools executes successfully , but result layer does not contains any output features..??
everything is set up correctly?
Move the post to
since it isn't an arcgis pro question
Did you try with existing pre-packed data?
I have tried it on the ESRI World Imagery Service in Woodland, California - exact same spot as shown in this Story Map . I am using ArcGIS Pro 2.7.2 with standard Deep Learning Framework installed. My deep learning setup is working perfectly as I have trained and detected numerous different MaskRCNN models. The model ran for 22 minutes and no footprints were returned. I even lowered the threshold to 0.2. I downloaded the ESRI pretrained model today. Parameters and settings to reproduce are in the attached screenshot.
I also tried it on our own 10cm aerial photography and footprints are returned ok. Not sure why the world imagery service doesn't work.
I have attached an image of the footprints I detected in the northwest corner of that image. If you can detect footprints with this image then that would suggest that your own imagery might not be usable with the ESRI model.