This Repository is used for our Road Sign Detection implementation in ROS running on a TurleBot3 with a Raspberry Camera.
In Ansatz_1 we tried an Approach with Keras which was trained with the GTRSB Database.
In Ansatz_2 we used transfer learning from a pre-trained SSD MobileNet with a Dataset composed of Images from Google and images we made ourself. There you can find create_tfrec.txt where the training of the Roadsign Detection is described with the used commands.
In the anaconda_dependencies.txt you can find the commands to install the needed dependencies.
