Shearlet-Based Structure-Aware Filtering for Hyperspectral and LiDAR Data Classification

Abstract

The joint interpretation of hyperspectral images (HSIs) and light detection and ranging (LiDAR) data has developed rapidly in recent years due to continuously evolving image processing technology. Nowadays, most feature extraction methods are carried out by convolving the raw data with fixed-size filters, whereas the structural and texture information of objects in multiple scales cannot be sufficiently exploited. In this article, a shearlet-based structure-aware filtering approach, abbreviated as ShearSAF, is proposed for HSI and LiDAR feature extraction and classification. Specifically, superpixel-guided kernel principal component analysis (KPCA) is firstly adopted on raw HSIs to reduce the dimensions. Then, the KPCA-reduced HSI and LiDAR data are converted to the shearlet domain for texture and area feature extraction. In contrast, superpixel segmentation algorithm utilizes the raw HSI data to obtain the initial oversegmentation map. Subsequently, by utilizing a well-designed minimum merging cost that fully considers spectral (HSI and LiDAR data), texture, and area features, a region merging procedure is gradually conducted to produce a final merging map. Further, a scale map that locally indicates the filter size is achieved by calculating the edge distance. Finally, the KPCA-reduced HSI and LiDAR data are convolved with the locally adaptive filters for feature extraction, and a random forest (RF) classifier is thus adopted for classification. The effectiveness of our ShearSAF approach is verified on three real-world datasets, and the results show that the performance of ShearSAF can achieve an accuracy higher than that of comparison methods when exploiting small-size training sample problems. The codes of this work will be available at http://jiasen.tech/papers/ for the sake of reproducibility.

CircWaveNet: A New Conventional Neural Network Based on Combination of Circlets and Wavelets for Macular OCT Classification

“…Therefore, in this study several geometrical X-let transforms including 2D-DWT (11, 25, 26) (Note that Haar wavelet was taken in current study), DTCWT (14, 27) (Note that just the real parts of this transform are utilized in this research to reduce the complexity and redundancy), shearlets (28, 29), contourlets (30), circlets (31), and ellipselets (24) were applied to decompose each B-scan a linear combination of basis functions or dictionary atoms. The non-subsampled (NS) (32) form of the multi-scale X-let transforms was employed to build a multi-channel matrix for each B-scan using all the sub-bands in parallel.…”

A new convolutional neural network based on combination of circlets and wavelets for macular OCT classification

“…Therefore, in this study, several geometrical X-let transforms, including 2D discrete wavelet transform (2D-DWT) 14,32,33 (Note that Haar wavelet was used in the current study), dual tree complex wavelet transform (DTCWT) 20,34 (Note that just the real parts of this transform are utilized in this research to reduce complexity and redundancy), shearlets 35,36 , contourlets 37 , circlets 38 , and ellipselets 15 were applied to decompose each B-scan into a linear combination of basis functions or dictionary atoms. The details of this step are illustrated in Fig.…”

Mapping and spatiotemporal dynamics of land-use and land-cover change based on the Google Earth Engine cloud platform from Landsat imagery: A case study of Zhoushan Island, China

“…Alijani et al studied the evolution characteristics of land-use/cover changes in Iran over 20 years, from 1996 to 2016 [ 23 ]. The data sources used in previous studies include hyperspectral, light detection and ranging (Lidar), moderate resolution imaging spectroradiometer (MOD)IS data, and Landsat data [ [24] , [25] , [26] , [27] ]. Classification algorithms include the 2-D convolutional neural network, hybrid convolutional network, and RF [ [28] , [29] , [30] ].…”