Overview / Usage

In the greater Pittsburgh area, there have been record rainfall and landslides within 2018. Responders and Surveyors are exasperated to assess and determine the risk the site posses due to constant rain and time consuming risk assessments. 3D Reconstruction is a viable solution; however, due to expense laser scanners or high quality dslr cameras, drone equipment and vast technical knowledge needed to navigate and install various software dependencies making this is rather inaccessible. We provide a simple framework for that only requires images from a smart phone and a general purpose laptop. We hope to enable responders another tool to quickly assess and track changes of sites over time.

Methodology / Approach

We incorporate the latest advancements in Structure from Motion (COLMAP); Point Cloud Segmentation, Registration and Alignment (Open3D); containers ({Nvidia-}Docker); gpu-proccessing (Cuda); Python; etc

[PROVIDE MORE INFO ON FOUNDATIONS OF THESE COMPONETS GO INTO BUNDLE ADJUST ETC]

Technologies Used

Colmap,
(Nvidia-)Docker,
Open3D,
Cuda,
Cloud Compare,
Images from Smart Phone Cameras