May I present you a collision-avoidance-system using the DepthAI OAK-D

For additional information to the project check out the detailed paper I wrote. Please also check out the added code, dataset, trained models, PowerPoint presentation and additional files. https://drive.google.com/drive/folders/1P4aJD1bQrliQuO2XWLKK_LW5oQck8lYP?usp=share_link

Object Detection, Distance Messurement and Path Finding

1. Clone the repo

git clone https://github.com/Jojo942004/Collision-Avoidance-System

2. download the folders from google drive to get access to custom models trained on boat images, pretrained models of the yolov5 repo, custom boat dataset

https://drive.google.com/drive/folders/1P4aJD1bQrliQuO2XWLKK_LW5oQck8lYP?usp=share_link

3. connect your OAK-D by Luxonis
4. simply run

python collision-avoidance-system.py

To gain access to trial data, custom models please visit
https://drive.google.com/drive/folders/1P4aJD1bQrliQuO2XWLKK_LW5oQck8lYP?usp=share_link

Presentation Collision Avoidance System.pdf

depthai==2.17.4.0
torch==1.12.0+cu116
torchaudio==0.12.0+cu116
torchvision==0.13.0+cu116
opencv-contrib-python==4.6.0.66
opencv-python==4.6.0.66
numpy
math
multiprocessing
time

Johannes Zimmermann @Jojo942004