Aircraft Type Recognition in High-Resolution SAR Images Using Multi-Scale Autoconvolution

Li Ping Zhang; Chao Wang; Hong Zhang; Bo Zhang

doi:10.4028/www.scientific.net/KEM.439-440.1475

Paper Titles

Study and Analysis of Shock Wave Propagation in Excavation Laneway during Methane Explosion
p.1450

In Vitro Selection of Oligonucleotide Acid Aptamers against Pathogenic Vibrio Alginolyticus by SELEX
p.1456

A Novel VoIP Scheduling Algorithm in IEEE 802.16e System
p.1463

Measurement of Worker Differences and Optimization of Personnel Assignment for Assembly Production
p.1469

Aircraft Type Recognition in High-Resolution SAR Images Using Multi-Scale Autoconvolution
p.1475

A Improved Topics Search Algorithm Based on PSO Strategy for Web Mining
p.1481

The Scheduling Algorithm of Grid Task Based on PSO and Cloud Model
p.1487

A Novel Binary Differential Evolution for Discrete Optimization
p.1493

A Supporting Framework for Real-Time Data Mining
p.1499

HomeKey Engineering MaterialsKey Engineering Materials Vols. 439-440Aircraft Type Recognition in High-Resolution SAR...

Aircraft Type Recognition in High-Resolution SAR Images Using Multi-Scale Autoconvolution

Abstract:

Automatic target recognition is the key stage of SAR image interpretation system and has been taking a great interest to the researchers in recent years. Aiming at the issue of aircraft type recognition in high-resolution SAR images, a novel method based on multi-scale autoconvolution (MSA) affine invariant moment is proposed. First, the texture analysis and clustering method are used to segment the SAR images and then the denoising algorithm and morphological processing are applied to segmented results. Second, 29 MSA features are extracted and form a feature vector to represent the target, then the vector components are standardized by gauss normalization. In the final, the vectors are classified by using the nearest neighbor classifier and template library constructed previously. Experimental results show that the proposed method can obtain high accuracy rate with high processing speed, in which the accuracy rate of two type aircrafts with real data arrives at 85.17% and the accuracy rate of four type aircrafts with simulated data arrives at 87.85%.