Abstract
Hyperspectral images are rich in both spectral information and spatial dependence information between pixels; however, hyperspectral images are characterized by the high dimensionality of small data sets and the spectral variance. Facing these problems, spatial dependence information as supplementary information is a relatively effective means to solve them. And the label dependence characteristic of hyperspectral images is excellent spatial dependence information. Therefore, to address the above issues, based on residual network and spatial information extractor(RAS), which is based on a residual network, pixel embedding(PE), and a spatial information extractor(SIE). At the stage of mining spectral information, we use the residual network to mine spectral features; At the stage of mining spatial information, we utilize the label dependency characteristic to feed the set of pixels containing the target pixels into PE. Then, a pixel vector with location information and self-defined dimensionality is obtained. Next, this vector is fed into our proposed SIE to mine the spatial dependency information. In multi-group ablation experiments, our proposed model achieves overall accuracy (OA) scores of 79.16% on the 5% Indian Pines test set, 90.82% on the 1% Pavia University test set, and 92.17% on the 1% Salinas test set. Especially, the experimental results demonstrate that the joint spectral-spatial approach is effective in improving the accuracy of hyperspectral image classification.
Highlights
Hyperspectral images have continuous, multiband narrow spectral bands [1]
Hyperspectral image classification is the classification of pixel points in an image, and the usual method is to use a priori information in the image, such as a small number of labeled training samples, to learn to discriminate the classes corresponding to other pixels in the hyperspectral image
We exploit the label dependence property in hyperspectral images and use the surrounding pixels of pixels to be classified as spatial information
Summary
Hyperspectral images ( known as remote sensing data) have continuous, multiband narrow spectral bands [1]. The wide spectral range carries substantial spatial and spectral information [2]. Due to the abundant information in hyperspectral images, the technology is used in many fields, such as agriculture [3], medicine [4], and food safety [5]. The recognition and classification of target objects in hyperspectral images have become an important direction for research in the hyperspectral image field [6]. Hyperspectral image classification is the classification of pixel points in an image, and the usual method is to use a priori information in the image, such as a small number of labeled training samples, to learn to discriminate the classes corresponding to other pixels in the hyperspectral image. Because the spectral range of hyperspectral images is wider than that of ordinary images, they carry more useful information in the continuous
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call