The Automotive Drone Analysis Tool (short: ADAT) is a Windows-only, desktop-based application that was developed as a submition for the Code Competition held by IAV GmbH. The application reached first place and therefore won the competition alongside its price. (See here)
- C#
- .NET Stack / .NET Core 3.0
- WPF => MVVM
- Visual Studio 2019
- Git
- Material Design
- YOLO (You-Only-Look-Once) Real-Time Object Detection
- Open Source Nerual Network Darknet
ADAT is a program that takes in drone images of certain traffic scenarios (e.g. parking cars) and detects all objects in them. The user may then adjust the drawn bounding boxes of these objects and adds a number of additional information (such as length of certain lines) to the images. After having prepared the image, the user can then analyse the picture in the dynamic report view, in which he can see distances between objects, angles to each other and lengths of each drawn line. Please refer to "Usage" for a quick example.
To start the analysis, the user may first choose between loading in images or video files.
After having chosen which images to analyse, YOLO tries to detect all objects in them. It may look like this:
The user then has to add more information to the already found objects to guarantee a meaningful analysis. A step by step guide can be found on the right side of the application named "Checklist". A fully prepared image may then look like this:
From there, the user may enter the Dynamic Report View, which should look like this after correctly preparing the image:
Every YOLO object is selectable and can be enabled or disabled, causing all reference lines and values of this object to also be visible or non-visible. That way, the user can choose specifically which objects he would like to analyse without the screen being convoluted by too much information. This analysis can also be exported to .pdf as seen here:
ADAT can also analyse videos of e.g. parking cars using the same formular. Via a small built-in video player, the user can then watch the analysed video. You can find an example below. (The first image leads to a Youtube video of the analysis.)
Thank you for visiting this project and also thanks to IT-Talents and IAV GmbH for hosting the competition! If you have any questions regarding the application or would like to use it, email me: [email protected].