SAR image target identification method based on target area and shadow joint decision

Abstract

The invention discloses a target area and shadow joint decision-making SAR image target identification method, and provides a target area and shadow joint SAR target identification method, which comprises the following steps of: describing boundaries of a target area and a shadow by adopting an elliptical Fourier descriptor; and describing a relative relationship between the two by using a characteristic vector. A sparse representation classifier is adopted to classify the boundary of the target area, the elliptic Fourier descriptor of the shadow and the characteristic vector of the relative relation, and a decision-making layer fusion method is adopted to obtain a more robust recognition result.

Description

technical field [0001] The invention relates to a SAR image target recognition method for joint decision-making of target areas and shadows. Background technique [0002] Synthetic Aperture Radar (SAR) target recognition is one of the key means of battlefield intelligence acquisition. Therefore, domestic and foreign scholars have carried out extensive and in-depth research on this field. SAR target recognition methods generally include two key steps, namely, feature extraction and classifier. The former obtains low-dimensional feature vectors from original SAR images through mathematical means or image analysis. Commonly used SAR image features include geometric shape features, projection features and scattering center features. Geometric features mainly describe the physical shape information of the target, typically the target area, outline, etc. The projection features seek the invariant features of the original high-dimensional SAR image in the low-dimensional subspa...

AI Technical Summary

Problems solved by technology

Most of the existing SAR target recognition methods directly use the overall SAR image or the segmented target area, while ignoring the role of SAR image shadows.

Images