A new object-based method is developed to extract moving vehicles and then automatically detect their velocities from two consecutive images. Several global and local values of values and gray sizes are examined to classify the objects in the image. Vehicles and their associated shadows can be discriminated against by removing large objects, such as roads. To detect speed, vehicles and shadows are first extracted from two consecutive images. The corresponding vehicles of the two images are then linked based on the similarity of shape and size and the distance within a threshold. Finally, using the distance between the corresponding vehicles and the delay between two images, we can detect the speed of movement and the azimuth angle. Our test shows promising results for detecting vehicle speeds. Later development will employ the proposed method for a pair of panchromatic and multi-spectral images of QuickBird, which have a thicker spatial resolution.
Satellite remote sensing and aerial photography have been used to capture snapshots of the Earth's surface. But if you take two images with a short time interval, they can be used to detect moving objects, eg. Cars, and even to measure their speeds. If this data is available, many new applications can be considered, such as observing traffic conditions on the roads. From the images taken by high-resolution satellites, p. QuickBird, we can observe small objects such as cars. QuickBird's panchromatic and multi-spectral sensors are known to have a slight time delay, approximately 0.2 seconds, depending on the scanning mode of the instrument. Using this time interval between two sensors of a scene, the speed of moving objects can be detected (Etaya et al., 2004; Xiong and Zhang, 2006). In this study, using the pansharpened images of airports, trains and freeways of Google Earth, the effect of the small time delay is demonstrated. The speed of operation of cars on a road with a QuickBird scene (a product of panchromatic and multispectral bands) from Bangkok, Thailand, is then evaluated as an example. Another method to detect the speed of the car is proposed using aerial photographs. Aerial photographs are often taken along a flight line with an overlap between adjacent scenes. If a moving object captured in a scene is also captured in an adjacent image, the speed of the object can be detected. A new object-based method is developed to extract moving vehicles and subsequently automatically detect their velocities from two consecutive aerial images. Using a pair of digital aerial images of central Tokyo, the proposed automatic extraction method is applied and the accuracy compared with the results of visual extraction is examined.