StarsMyDestination / ML-Car-Detection

Two Methods: 1.Classic computer vision 2.SSD-Single Shot MultiBox Detector

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Car detection and tracking

dataset

dataset-small dataset-big

Classic computer vision

featrues inclue:

  1. spatial bining color (raw pixels => color & shape)
  2. histogram of color (=>color)
  3. histogram of oriented gradients(hog) (=>shape)

drawbacks:

  1. computational expensive => low accuracy with a relative high speed (2fps)

SSD - Single Shot

About

Two Methods: 1.Classic computer vision 2.SSD-Single Shot MultiBox Detector


Languages

Language:Jupyter Notebook 89.7%Language:OpenEdge ABL 10.1%Language:Python 0.3%