Vehicle Extraction and Speed Detection from Digital Aerial Images

A new object-based method is developed to extract the moving vehicles and subsequently detect their speeds from two consecutive digital aerial images automatically. Several parameters of gray values and sizes are examined to classify the objects in the image. The vehicles and their associated shadows can be discriminated by removing big objects such as roads. To detect the speed, firstly the vehicles and shadows are extracted from the two images. The corresponding vehicles from these images are linked based on the order, size, and their distance within a threshold. Finally, using the distance between the corresponding vehicles and the time lag between the two images, the moving speed can be detected. Our test shows a promising result of detecting the moving vehicles' speeds. Further development will employ the proposed method for a pair of QuickBird panchromatic and multi-spectral images, which are at a coarser spatial resolution.