Using BRAILS for Computer Vision Applications in Natural Hazards Engineering

The latest advancements in artificial intelligence (AI) can be used to expedite tasks critical to natural hazards engineering. These tasks range from automated extraction of infrastructure characteristics from aerial and street-level imagery to crack-level damage detection from video recordings. This workshop covers many aspects of computer vision needed to understand the topic, and it aims to provide the fundamentals of various AI approaches applicable to existing problems in natural hazards engineering. These problems include modeling the built environment at a regional scale, rapid damage detection from reconnaissance data, and vision-based infrastructure monitoring. The workshop will utilize BRAILS, an AI-enabled tool that uses these methods for predicting critical attributes of the built environment from public imagery and automated detection of damage from reconnaissance data.

Prerequisites: Some Python programming knowledge is required. Participants w ith no or little Python experience should review the Python quick start material (Chapters 1-6) from SimCenter Programming Bootcamp 2024

Contents:

• Presenters
• Syllabus
  • Day 1: Image classification, detection, segmentation
  • Day 2: Vision foundation models
  • Day 3: Using BRAILS for inventory generation and damage detection
• Notebooks
  • Day 1: Image classification, detection, segmentation
  • Day 2: Vision foundation models
  • Day 3: Using BRAILS for inventory generation and damage detection

Note

1. Course material is available through the NHERI-SimCenter/SimCenter_BRAILS_WORKSHOP_2024 repository on GitHub

2. You will be using Jupyter Notebooks on DesignSafe to perform the exercises.