P4 - Automatic Inspection of sewers

The sampled code for Group 464, P4

Description

This is a coding project for AAU robotics fourth semester. The project tries to detect sewer damages, in a simulated setup. The repository includes files for detection of damages, training data generation, and files for using the Microsoft Kinect 2's depth and image data.

Dependencies

This code requires:

  • A Microsoft kinect 2 for windows (Or videos recorded through one)
  • Python 3.8
  • openCV 4.0.1
  • matplotlib 3.5.1
  • numpy 1.22.3
  • nxt 0.1
  • pykinect2 0.1.0
  • scikit_learn 1.1.1

Navigating the folders

Main files for driving the NXT, recording with the kinect, and damage detection, is found front of the folders.

The \data folder contains .pkl files, such as trained classifiers and test data. (Trained classifiers are not uploaded to git due to their file size)

The \Functions folder contains files that are referenced in others, such as classes and functions.

The \training_data_files folder contains files for generating training data.

The \tests folder contains various programs developed to test the functionality of software developed throughout the project.

Authors

Contributors names and contact info

Miki Barlach Pregaard mprega20@student.aau.dk

Wallat Bilal wbilal18@student.aau.dk

Frederik Saldern Nielsen fsni20@student.aau.dk

Glenn Gadensgaard Svendsen gsvend20@student.aau.dk

Simon Haaning Andersen san20@student.aau.dk

Asbjørn Lauritsen alau20@student.aau.dk