This repository contains Deep Learning Crack Detection Algorithm for the conference paper "Automated Building Exterior Crack Inspection using UAVs, Open-Source Deep Learning and Photogrammetry"
Before Running, Install the following
pip install tensorflow keras matplotlib image numpy Pillow
and execute the following if you are using Jupyter Notebook on Ubuntu
source my_project_env/bin/activate
Place the image to be analyzed in the input folder.
Open Concrete Wall Crack Detection.ipynb file, modify the input image name and execute the final code block.
The Webot Simulation file used in the paper is in Webot Simulation folder. Running this simulation on Ubuntu is recommended. The controller is written in C.
For training with own data set, have a folder at root named Data