This is the project repo for the final project of the Udacity Self-Driving Car Nanodegree: Programming a Real Self-Driving Car. For more information about the project, see the project introduction here.
Team Lead and Member
Kan-Hua Lee/ kanhwa@gmail.com (single member)
Installation and Requirements
Detailed installation instruction to run this project is described here
This project requires
- Python 2.7.12
- Tensorflow 1.3.0
- OpenCV 3.3.1
Notes on Traffic light Classification Model
My classification model requires a trained MobileNet model. It will be downloaded automatically when first running this project.
If somehow the model cannot be downloaded automatically, you can download it from this link and
extract it to folder tl_detector/.
Since my VM does not have a GPU to run Tensorflow,
I need to reduce the number of images to be classified to not to drag the car.
The default setting only run 1/20 of the images sending to the tl_detector.
This can be altered by changing DETECT_FREQ in tl_detector.py.