A Label Similarity Probability Filter for Hyperspectral Image Postclassification

Abstract

This paper presents a Label Similarity Probability Filter (LSPF) to make hyperspectral image post-classification. The LSPF is inspired from the first law of geography and proposes a class label probability function to quantify the probability of both centered and its neighboring pixels in belonging to the same class. It first classifies the hyperspectral data using the regular support vector machine classifier. Then, it binarizes the posterior classification result to obtain the binary label maps of each class. After that, it traverses all spatial windows centered by each pixel and calculates the cumulative probability of all pixels in each class. Finally, the cumulative probabilities are used to make reclassification to obtain the refined classification map. Experiments on Indian Pines, Pavia University and ZY1-02D Yellow River Estuary data show that LSPF greatly improves the classification accuracy of spectral signatures and outperforms other state-ofthe-art spectral-spatial methods IndexTerms-hyperspectral; post-classification; label similarity probability filter; spectral-spatial methods.

Mapping Coastal Wetlands Using Transformer in Transformer Deep Network on China ZY1-02D Hyperspectral Satellite Images

“…The ZY1-02D satellite, launched on September 12, 2019, is the first self-built commercial hyperspectral satellite in China [42]. It can be utilized to large-scale observation and quantitative remote sensing missions with high spectral resolution and medium spatial resolution.…”