Deep learning model to extract building footprints from high-resolution aerial and satellite imagery. A brief summary of the item is not available. Add a brief summary about the item.
Deep learning package
from
Item created: Sep 29, 2020 Item updated: Dec 30, 2024 Number of downloads: 109,626
Description
Building footprint layers are useful in preparing base maps and analysis workflows for urban planning and development. They also have use in insurance, taxation, change detection, infrastructure planning, and a variety of other applications.
Digitizing building footprints from imagery is a time-consuming task and is commonly done by digitizing features manually. Deep learning models are highly capable of learning these complex semantics and can produce superior results. Use this deep learning model to automate the tedious manual process of extracting building footprints, reducing time and effort required significantly.
Using the model
Follow the guide to use the model. Before using this model, ensure that the supported deep learning libraries are installed. For more details, check Deep Learning Libraries Installer for ArcGIS.This model can be fine-tuned using the Train Deep Learning Model tool. Follow the guide to fine-tune this model.
Input
8-bit, 3-band high-resolution (10–40 cm) imagery.Sample results
An in-depth description of the item is not available.
Layers
Tools
Tables
Basemap
Project Contents:
Solution Contents
Contents
Layers
Screenshots
Terms of Use
No special restrictions or limitations on using the item's content have been provided.
Details
Dashboard views: Desktop
Source:
Creating data in:
Published as:
Other Views:
Dependent items in the recycle bin
Applicable: 2d
Size: 157.818 MB
ID: a6857359a1cd44839781a4f113cd5934
Image Count: 0
Image Properties
Layer Drawing
Using tiles from a cache
Dynamically from data
Share
Owner
Folder
Categories
This item has not been categorized.
Credits (Attribution)
No acknowledgements.Esri, Maxar
Comments (66)
I've run this model on a few sample areas and it works great for single dwelling homes but the model really seems to struggle with larger commercial/industrial buildings. I have 1" imagery that I resampled down to 10 cm before running the model. What parameters can I adjust so that I get better results on larger buildings? Here's the arguments I'm using to run the model: Padding: 128 batch_size: 16 threshold: 0.9 return_bboxes: False test_time_augmentation: False merge_policy: mean tile_size: 512 Any suggestions?
Great to hear! To detect larger buildings, try increasing the tile_size argument to 1024.
can this BFE be used for industrial/ power generation sites or just residential buildings? Seeking to understand the range of applicability of this module?
This model is trained on U.S. geography and is designed to detect buildings within this geography including industrial or power generation sites.
@spathak_deldev Thank you so much for clarifying that. Whole time I was using the Extract Features Using AI Models tool... as I obtained some bad information from a video on it. The Detect Objects Using Deep Learning tool did exactly what I was hoping, thanks so much for your help!
Hi @spathak_deldev thanks for your reply. I am trying to extract the Red, Green, and Blue Bands from my imagery, since the imagery has a 4th band for Infrared I believe, and this model only takes 3 - so I assumed the RGB bands are most necessary. I have gotten the Classify Pixels Using Deep Learning tool to work for a different model (Parcel Extraction), however I prefer the idea of a tool exporting features instead of another raster, specifically individual houses. And so, I have used the extract bands raster function, per your suggestion, to extract the RGB bands that way, and the model still has the same error as prior. I will try with a different combination of bands.
Hi @JAllan, can yoj please comfirm which model you are trying to use. If you wants to extract building footrints using Building Footprint Extracrtion - USA model you will have to use Detect Objects Using Deep Learning tool.
Hello - I am trying to use this model on some USGS imagery I downloaded. However, those rasters from the imagery contain 4 bands, and this model only runs on rasters with 3 bands. When I try to extract the raster with 4 bands to a new raster with 3 bands (RGB), the tool now gives me an error "The value is not a member of HIGH RESOLUTION LAND COVER CLASSIFICATION - USA". Any suggestions? It seems whenever I try to extract 3 bands from the 4-band raster, its no longer classified as High Resolution, but then the tool won't work on 4-band rasters. Any help appreciated
Hi, Use can use Extract Bands raster function and use that raster as input in the Classify Pixels Using Deep Learning tool. Which bands are you extracting?