Shape and color features based airport runway detection

In this paper, a novel, simple and efficient airport runway detection algorithm is proposed. Proposed algorithm consists of two stages. First stage of the algorithm coarsely segments the runway. This stage uses anisotropic diffusion and Frangi filter. Anisotropic diffusion provides the noise immunity and Frangi filter detects all possible runway candidates. To avoid the false candidates generated in this first stage, the candidate pixels are passed through a second stage using shape and chroma based features. Qualitative and quantitative analysis performed on various images show that proposed method can detect runways effectively for both simple and complex scenarios. Our proposed algorithm has a significant role for both defence and commercial applications.