研究目的
To propose an effective method for hyperspectral image super-resolution that considers the local geometrical structure of the sparse coefficients, incorporating their local similarity into the sparse coding process to preserve the properties of local geometrical structures.
研究成果
The proposed method demonstrates superiority over other methods in reconstructing high-resolution hyperspectral images by incorporating the local geometrical structure of the sparse coefficients into the sparse coding process.
研究不足
Not explicitly mentioned in the abstract.
1：Experimental Design and Method Selection:
The method exploits the location related constraint about the sparse coefficients and incorporates their local similarity into the sparse coding process.
2：Sample Selection and Data Sources:
Experiments were conducted on two popular datasets of hyperspectral images: Indian Pines and Pavia Center.
3：List of Experimental Equipment and Materials:
Not explicitly mentioned.
4：Experimental Procedures and Operational Workflow:
The ground truth image is downsampled by averaging over disjoint 32×32 blocks to obtain the low spatial resolution HSI X. For MSI Y, several bands are directly selected from the ground truth image.
5：Data Analysis Methods:
Three widely used metrics: root mean square error (RMSE), relative dimensionless global error in synthesis (ERGAS) and spectral angle mapper (SAM) are employed to evaluate the quality of reconstructed hyperspectral images by different methods.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
