Carnd-Vehicle-Detection-YOLOv3-darknet-

This is the yolov3 darknet implementation of fifth project of Udacity term-1 carnd nano-degree program

OVERVIEW

Uses the lastest version of yolo in darknet to completed the project.

REQUIREMENTS

Pre-install the environment of carnd nanaodegree program( mainly moviepy and cv for this code)

INSTALLING DARKNET

If you have both

NUMPY in -I/usr/include/python2.7/ -I/usr/lib/python2.7/dist-packages/numpy/core/include/numpy/

CUDA in -I/usr/local/cuda/include/

present in the above location in your gpu system u can use the libdarknet.so already present in the folder

If you dont have a GPU system or numpy and cuda in specified locations, follow the below steps

1. Download the file by git clone https://github.com/pjreddie/darknet
2. Implement the changes from the issue in pjreddie/darknet#289 to load the image using moviepy or numpy and predict
3. make the code
4. Now a new libdarknet.so will be present in the folder copy that and replace in this folder

EXECUTION OF THE CODE

python abstract.py input_video_file_name output_video_file_name