Monocular Vision for Distance Estimation Using Deep Learning.
The following is a Deep Learning distance estimator.
A CNN is used to predict the distances of pedestrians in a live video.
The video is recorded using monocular vision (Single camera).
Instead of dual cameras (Stereo vision) as used in industry for distance estimation.
Processes label file info from the KITTI dataset.
The pedestrian's info is then stored locally.
The trained CNN is imported.
A video of pedestrians is loaded.
The predicted distance of the pedestrians is displayed on the output video.