Abstract:The algebraic multigrid (AMG) method is used to solve linear systems of equations on a series of progressively coarser grids and has recently attracted significant attention for image segmentation due to its high efficiency and robustness. In this paper, a novel spectral-spatial classification method for hyperspectral images based on the AMG method and hierarchical segmentation (HSEG) algorithm is proposed. Our method consists of the following steps. First, the AMG method is applied to hyperspectral imagery to construct a multigrid structure of fine-to-coarse grids based on the anisotropic diffusion partial differential equation (PDE). The vertices in the multigrid structure are then considered as the initial seeds (markers) for growing regions and are clustered to obtain a sequence of segmentation results. In the next step, a maximum vote decision rule is employed to combine the pixel-wise classification map and the segmentation maps. Finally, a final classification map is produced by choosing the optimal grid level to extract representative spectra. Experiments based on three different types of real hyperspectral datasets with different resolutions and contexts demonstrate that our method can obtain 3.84%–13.81% higher overall accuracies than the SVM classifier. The performance of our method was further compared to several marker-based spectral-spatial classification methods using objective quantitative measures and a visual qualitative evaluation.

Segmentation of Hyperspectral Remote Sensing Image Based on Morphological Gradient Reconstruction and Marker Extraction

Semantic image segmentation integrating gradient reconstruction and non-linear processing

Image Segmentation Based on Watershed Transform

Time-Frequency Spectrogram Segmentation Using the Multi-Scale Morphological Gradient and the Marked Watershed Algorithm

IMAGE SEGMENTATION: A WATERSHED TRANSFORMATION ALGORITHM

A Marker-Based Watershed Segmentation for High Resolution Remote Sensing Image

Multispectral Image Segmentation by A Multichannel Watershed-Based Approach

A method of image reconstruction for easy segmentation

Segmentation of High Resolution Remote Sensing Image Via an Improved Marker-Based Watershed Algorithm

A Spectral-Spatial Classification of Hyperspectral Images Based on the Algebraic Multigrid Method and Hierarchical Segmentation Algorithm

Mathematical Morphological Segmentation of IKONOS Multispectral Data

Evaluation of Multiscale Morphological Segmentation of Multispectral Imagery for Land Cover Classification.

A Hierarchical Segmentation Method for Multispectral Imagery

An Unsupervised Marker Image Generation Method for Watershed Segmentation of Multispectral Imagery

Watershed Distribution Equipment Infrared Thermal Image Segmentation Method Based on the Morphological Marker

Multispectral Remotely Sensed Imagery Segmentation Based on First Fundamental Form

Building Extraction from Remote Sensing Image Based on Improved Marker-controlled Watershed Algorithm

An Algorithm of Multi-scale Image Segmentation Based on Morphology

Image Segment Algorithm Based on Adaptive Image Preprocessing

Fast Segmentation Algorithm of High Resolution Remote Sensing Image