Shidhani/birds-eye-view

Obtain bird's eye view of a scene from a single input image

birds-eye-view

This repo is a copy of the original repo, which is an implementation of the paper: "A Geometric Approach to Obtain a Bird's Eye View from an Image", Ammar Abbas, Andrew Zisserman. The code estimates the homography matrix for a bird's eye view transformation along with some of the cameras's intrinsic and extrinsic parameters.

How to use:

• To run the code, you need to use an environment with TensorFlow 1.8 or 1.13.2, OpenCV, Numpy, and Matplotlib.

• Make sure the following file are installed completely:

  data/saved_models/incp4/model.ckpt-17721.data-00000-of-00001

  data/saved_models/vgg16/model.ckpt-20227.data-00000-of-00001

  the files can be found here.

• To predict bird's eye view from an image or a video, run the Following command:

  python scripts/predict_horizon_vpz_homography.py --view_path $VIEW_PATH --model_name $CNN_MODEL

  Where:

  $VIEW_PATH: path to the image/video

  $CNN_MODEL: name of the cnn model (available options: vgg-16, inception-v4)

  • Sample images and videos are provided in images and videos folders.
  • The output of the homography matrix will be saved as a .txt file in output folder along with the transformed image/video.