Synthetic vision assisted real-time runway detection for infrared aerial images

Abstract

This paper presents a new real-time runway detection based on synthetic vision and level set method. It mainly focuses on the initial level set function and time performance. As for the initial level set function, three-thresholding segmentation is derived to obtain the subset of the runway, which serves as an initial curve to induce the initial level set function. As for time performance, a ROI (Region of Interest) based evolution method is proposed. Analysis of experimental results and comparisons with existing algorithms demonstrate the efficiency and accuracy of the proposed method.