Transactions of Nanjing University of Aeronautics & Astronautics
Super-Resolution Inversion and Reconstruction of Remote Sensing Image of Unknown Infrared Band of Interest
- Article
- Figures
- Metrics
- Preview PDF
- Reference
- Related
- Cited by
- Materials
Abstract:
This paper proposes a super-resolution inversion and reconstruction algorithm for remote sensing images of unknown bands of interest. The proposed method utilizes a built-in spectral reflectance database and the existing multi-spectral image to achieve the accurate classification of substances through the implementation of a Gaussian hybrid clustering algorithm and correlation distance method after radiation calibration and atmospheric correction. The image mixing algorithm based on the manifold space constraint obtains the distribution of ground substances， based on which the spectral reflectance image of the unknown band of interest is reconstructed. By employing the single-window algorithm， the temperature field of low-resolution ground substances is inverted through the far-infrared image， and the Shepard interpolation algorithm is used to interpolate the low-resolution temperature field to obtain a high-resolution ground temperature field. According to the spectral reflectance and the temperature field of the ground substances， using the remote sensing link imaging model， the high-resolution remote-sensing image of unknown infrared band of interest is super-resolution inversion reconstructed. Experimental results show that the reconstructed infrared image of various scenes has a high similarity with the original scene image， which has great benefits for improving the ability of target detection and recognition.
Keywords:
Project Supported:
This work was supported by the National Defense Science and Technology Foundation Strengthening Program （No.2021-JCJQ-JJ-0834）， and the Fundamental Research Funds for the Central Universities （Nos.NJ2022025， NP2022450）.
Clc Number:
TP751.1
Mobile website
