Herramientas/Tool Index
Unetseg
What is it?
It is a Python package, a deep learning model for semantic segmentation based on the U-Net architecture. It is suitable for any use case where you want to delimit the outline of an object of interest, and it is specially tuned for multispectral satellite, aerial and drone images of different spatial resolution.
Why was it developed?
To facilitate the task of preparing the data and being able to train a machine learning model that can capture satellite images of different areas through drones of different spatial resolution.
Why are we sharing it?
The purpose of the same is to share the development of the code in an open way, which allows other institutions, entities or cities, to modify, study or provide information in addition to having the source code at their disposal.
Have you reused or are you thinking of reusing this tool?
Enter our competition to win several prizes, including a trip to participate in the Developer Week, February 15-23 of 2023, in San Francisco, CA – US.
Email us at code@iadb.org and request an exclusive “tool reuse badge” from the IDB.
Have you participated in the implementation workshop of this tool?
The attendees (live or through our YouTube channel) who send the results of the exercises of our workshops to the email code@iadb.org will receive the badge by email within a week. To request a new credential submission, please contact: code@iadb.org.
