A multilateral filtering method applied to airplane runway image

TL;DR

This paper introduces a multilateral filtering method for airport runway images that enhances noise removal while preserving edges, outperforming standard bilateral filtering through steerable filtering and texture analysis.

Contribution

The paper presents an improved multilateral filtering technique that combines steerable filtering and texture features for better noise reduction in runway images.

Findings

Multilateral filtering outperforms standard bilateral filtering.

Texture features improve filtering effectiveness.

Simulation results confirm enhanced noise removal.

Abstract

By considering the features of the airport runway image filtering, an improved bilateral filtering method was proposed which can remove noise with edge preserving. Firstly the steerable filtering decomposition is used to calculate the sub-band parameters of 4 orients, and the texture feature matrix is then obtained from the sub-band local median energy. The texture similar, the spatial closer and the color similar functions are used to filter the image.The effect of the weighting function parameters is qualitatively analyzed also. In contrast with the standard bilateral filter and the simulation results for the real airport runway image show that the multilateral filtering is more effective than the standard bilateral filtering.