Speed-Estimation-Computer-Vision-Project

Ensure you have download python ver > 3.8 !!!

Install CUDA if you have NVIDIA GPU

Step for run the code:

1. Clone repository

git clone https://github.com/AmeliaSyahla/Speed-Estimation-Computer-Vision-Project.git
cd Speed-Estimation-Computer-Vision-Project

2. Activate your virtual environment Windows:

   python -m venv venv
   venv\Scripts\activate

   Linux:

   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies

   pip install --upgrade pip
   pip install -r requirements.txt

4. Install Pytorch dengan CUDA

   pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

5. Run code in your teminal

   python filename.py

   Dont forget to save your code before run, or you will get some error !!!

Notes

1. You must download video dataset before you run all
2. File 01_extract_frames.py contains step to make Dataset/ format video to frame. It will use for speed estimation. When you run this, the code will make new folder 'frames'
3. File 02_annotate_calib.py contains step to make you annotating through point and line to calibration and homography. This code will make new folder source_materials
4. File 03_calib_homography.py contains step to connect a single point from the previous annotation. This code will make new file filename_lines in source_materials folder
5. You can use SSD Model or Faster-RCNN to detection object and tracking, when you want to use Faster-RCNN, just run 04_detectcnn.py. This code will automatically downloaded model .pth and make new folder Output_Json_Data and Output_Videos
6. To look the output videos, you can access it on your local folder