Deep learning model to delineate agricultural fields using Sentinel-2 imagery. A brief summary of the item is not available. Add a brief summary about the item.
Deep learning package
by
Item created: May 18, 2023 Item updated: Dec 30, 2024 Number of downloads: 3,974
Description
The
delineation of agricultural field boundaries has a wide range of
applications, such as for crop management, precision agriculture, land
use planning and crop insurance, etc. Manually digitizing agricultural
fields from imagery is labor-intensive and time-consuming. This deep
learning model automates the process of extracting agricultural field
boundaries from satellite imagery, thereby significantly reducing the
time and effort required. Its ability to adapt to varying crop types,
geographical regions, and imaging conditions makes it suitable for
large-scale operations.
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.
Fine-tuning the model
This model can be fine-tuned using the Train Deep Learning Model tool. Follow the guide to fine-tune this model.
Input
Sentinel-2
L2A 12-bands multispectral imagery using Bottom of Atmosphere (BOA)
reflectance product in the form of a raster, mosaic or image service.
Output
Feature class containing delineated agricultural fields.
Applicable geographies
The model is expected to work well in agricultural regions of USA.
Model architecture
This model uses the Mask R-CNN model architecture implemented in ArcGIS API for Python.
Accuracy metrics
This model has an average precision score of 0.64 for fields.
Training data
This model has been trained on an Esri proprietary agricultural field delineation dataset.
Limitations
This
model works well only in areas having farmlands and may not give
satisfactory results in areas near water bodies and hilly regions. The
results of this pretrained model cannot be guaranteed against any other
variation of the Sentinel-2 data.
Sample results
Here are a few results from the model.
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: 169.49 MB
ID: eb5f896bf88b46af8252e17fa404a73d
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
Comments (5)
When can we expect a model that is compatiable with areas outside of the US?
When the geography of an area resembles that of regions in the USA where the model was trained, the model is likely to perform effectively. However, actual inferencing is necessary to confirm this, as the model's performance is influenced by various factors, including image resolution and the cell size specified in the tool.
We trained the model using Sentinel-2 Cloud-Optimized GeoTIFFs from Registry of Open Data on AWS: https://registry.opendata.aws/sentinel-2-l2a-cogs/ Since the bands are separate .tif files and of varying spatial resolutions we used a Mosaic Dataset to manage everything. You can leverage this public repository to automate creation of your own Mosaic Datasets to use with this Deep Learning Model. https://github.com/dkwright/arcgis-sentinel-2-cog-ag-fields/ . There is an example you can run from the batchfiles folder that will create the Mosaic Dataset for San Luis Valley for the month of July in 2023 using a 20% cloud cover limit and it adds a couple dozen processing templates for rendering various imagery composites and band indices on-the-fly. The batchfile can be modified to use your own Area of Interest, acquisition date range, and cloud cover percentage threshold. No download of imagery is required as the COG files are simply referenced by the Mosaic Dataset. There is an Esri Community thread here for additional conversation: https://community.esri.com/t5/arcgis-image-analyst-questions/agriculture-field-delineation-dlpk/m-p/1355441#M529
How can I download the Sentinel-2 Level2A Imagery?
For the input data, which Sentinel bands should be included? The documentation currently says use a 15 band Sentinel image at 10m resolution. Sentinel-2 has only 12 multispectral bands, 4 of which are 10m resolution. Should we also include some of the additional quaily or cloud bands? Please clarify, thank you!