Runway Positioning and Moving Object Detection Prior to Landing

Autor: M Ibrahim Mohamed, Rida M. Hamza, Dinesh Ramegowda, Venkatagiri Subbaraya Rao
Rok vydání: 2009
Předmět:
Zdroj: Augmented Vision Perception in Infrared ISBN: 9781848002760
DOI: 10.1007/978-1-84800-277-7_11
Popis: Safe navigation of both manned and unmanned aircraft requires a robust runway identification process and reliable obstacle detection to determine the runway status before landing. The navigation data extracted from multiple sources of current synthetic navigation sources are not adequate for positioning an aircraft, and it cannot detect moving obstacles on runways, especially during reduced visibility conditions. The enhanced vision system (EVS) described in this article can augment current synthetic vision database capabilities by providing more accurate positioning of a runway and detecting moving objects on it from an onboard infrared sensor. Our EVS is based on a two-step process. We first analyze the sensor image to identify and segment the runway coordinates. These estimates are then used to locate the runway structure and detect moving obstacles. In the segmentation process, we apply an adaptive thresholding technique to calculate the edges of a runway based on the predicted synthetic data. To match a runway template on the edges, we examine alternative fitting models. The predicted coordinates and the detected edges are then correlated to determine the location of the actual runway coordinates within the sensor image. These coordinate estimates are fed to the dynamic stabilization of the image sequence in the obstacle detection process. We also use feature points beyond the estimated runway coordinates to further improve the stabilization process. Next, we normalize the stabilized sequence to compensate for the global intensity variations caused by the gain control of the infrared sensor. We then create a background model to create an appearance model of the runway. Finally, we identify moving objects by comparing the image sequence with the background model. We have tested our EVS and reported significant improvements over the synthetic navigation data. We have been able to detect distant moving objects around the runway.
Databáze: OpenAIRE