Combining feature fusion and decision fusion for classification of hyperspectral and LiDAR data

Abstract

This paper proposes a method to combine feature fusion and decision fusion together for multi-sensor data classification. First, morphological features which contain elevation and spatial information, are generated on both LiDAR data and the first few principal components (PCs) of original hyperspectral (HS) image. We got the fused features by projecting the spectral (original HS image), spatial and elevation features onto a lower subspace through a graph-based feature fusion method. Then, we got four classification maps by using spectral features, spatial features, elevation features and the graph fused features individually as input of SVM classifier. The final classification map was obtained by fusing the four classification maps through the weighted majority voting. Experimental results on fusion of HS and LiDAR data from the 2013 IEEE GRSS Data Fusion Contest demonstrate effectiveness of the proposed method. Compared to the methods using single data source or only feature fusion, with the proposed method, overall classification accuracies were improved by 10% and 2%, respectively.

Characterizing Dryland Ecosystems Using Remote Sensing and Dynamic Global Vegetation Modeling

“…However, these methods are not always straightforward and may need more calibration to build a robust classification model. There is no doubt that integrated lidar-hyperspectral classification leads to better results (Asner et al, 2008;Koetz et al, 2008;Liao et al, 2014;Xu et al, 2018). The framework and approach described here are straightforward to implement and interpret, and are consistent with our ecological understanding.…”

SS-MAE: Spatial–Spectral Masked Autoencoder for Multisource Remote Sensing Image Classification

“…There are two types of supervised multi-source classification methods: decision-fusion [19] [20] [21] and featurefusion [22] [23] [24]. Specifically, decision-fusion methods usually use several classifiers separately and then combine the classifications together to form the final classification result.…”

“…Many decision-level fusion [19] [20] [21] and featurelevel fusion [22] [23] [24] methods have been devoted for multi-source remote sensing image classification, and achieved promising performance. However, these methods commonly employ supervised training strategy, which requires substantial amounts of labeled data, making it a labor-intensive process.…”

Object-oriented SVM classifier for ALSAT-2A high spatial resolution imagery: A case study of algiers urban area

“…Its objective is to use the data contained in remotely sensed images we are analysing to extract information and identify targets into a defined number of thematic classes (vegetation, soil, urban, forest, etc.). There are many different classification approaches [2][3][4]. They all fall under two main topics: unsupervised and supervised classification.…”