Car Detection in Low Resolution Aerial Image

In this paper, we present a system to detect passenger cars in aerial images where car appears small. We post the detection as a 3D object recognition problem to account for the variation in viewpoint and the shadow. We started from psychological test to nd important features human use to detect cars. Directed by the above result, we chose the boundary of the car body, the boundary of the front windshield, the the shadow as features. We observed some of these features are a ected by the intensity of the car and whether or not there is shadow along it. This information is represented in the structure of the Bayesian network which we use to integrate all features. Experiment shows very promising result even on some very challeging images.

• Type: Conference
• Year: 2001
• Where: ICCV
• Authors: Tao Zhao, Ramakant Nevatia
• Keywords: Car Body, Computer Vision, ICCV 2001, Important Features, Object Recognition Problem, Passenger Cars