Abstract:Understanding the scenes provided by remote sensing (RS) images has drawn increasing attention in the last decade. It is to automatically classify a RS scene image by feature extraction and label assignment. Although much effort has been dedicated to developing discriminative feature extraction as well as automatic classification techniques, it is still very challenging owing to the complex distributions of the ground objects in high spatial resolution scenes. To enhance the ability to represent the RS scenes, an integration of multiple types of features for remote sensing scenes is considered as an effective way. Nevertheless, different kinds of features possess different characteristics, and how to fuse them together is a critical problem. In this paper, to fuse three different but complementary types of features that are extracted to characterize global attributes of scenes together, a discriminant correlation analysis based feature-level fusion approach is proposed, which maximizes the correlation of corresponding features across the two feature sets and in addition de-correlates features that belong to different classes within each feature set. In addition, to further improve the scene classification performance, another kind of information fusion form, called decision-level fusion is adopted in the classification stage. In our proposed decision-level fusion technique, the final results of multiple classifiers are combined via majority voting. Extensive experiments results conducted on the well-known SIRI-WHU dataset, the WHU-RS dataset and the UC Merced Land Use dataset have demonstrated that the proposed remote sensing scene classification method based on heterogeneous feature extraction and multi-level fusion is superior to many state-of-the-art scene classification algorithms.

Heterogeneous Remote Sensing Image Fusion Based on Homogeneous Transformation and Target Enhancement

Infrared and Radar Fusion Detection Method Based on Heterogeneous Data Preprocessing

Unaligned Hyperspectral Image Fusion Via Registration and Interpolation Modeling

Fusion Of Infrared And Visible Light Images Based On Nonsubsampled Shearlet Transform

Homogeneous Transformation Based on Deep-Level Features in Heterogeneous Remote Sensing Images

The Realistic Fusion of Multi-spectral Images

Change Detection in Heterogeneous Optical and SAR Remote Sensing Images via Deep Homogeneous Feature Fusion

Hyperspectral and Multispectral Image Fusion Target Detection based on Cloud-Edge Collaboration

Change Detection in Heterogenous Remote Sensing Images via Homogeneous Pixel Transformation.

Deep-Learning-Based Spatio-Temporal-Spectral Integrated Fusion of Heterogeneous Remote Sensing Images

Multi-Spectral and Panchromatic Image Fusion Based on His-Wavelet Transform and Mopso Algorithm

An Integrated Framework for the Heterogeneous Spatio-Spectral-Temporal Fusion of Remote Sensing Images

Remote sensing image fusion algorithm based on à trous wavelet transform and HIS transform

Joint Feature-Space and Sample-Space Based Heterogeneous Feature Transfer Method for Object Recognition Using Remote Sensing Images with Different Spatial Resolutions

Change Detection in Heterogeneous Remote Sensing Images Based on the Fusion of Pixel Transformation.

Remote-sensing Image Fusion Based on IHS and Morphological Wavelet Transforms

Universal high spatial resolution hyperspectral imaging using hybrid-resolution image fusion

FTDN: Multispectral and Hyperspectral Image Fusion With Diverse Temporal Difference Spans

Heterogeneous image transformation

HFM: A hybrid fusion method for underwater image enhancement

Remote Sensing Scene Classification Using Heterogeneous Feature Extraction and Multi-Level Fusion