研究目的
To introduce a new spectral unmixing method for hyperspectral images based on secant function optimization that efficiently reduces spectral distortion.
研究成果
The proposed method (HUSFO) efficiently extracts the abundance maps by minimizing the secant error between the original hyperspectral images and the multiplication of endmember and abundance fraction matrices. It shows lower RMSE, AE, and AAD errors compared to other state-of-the-art linear unmixing methods.
研究不足
The main complexity of the proposed method is related to the SQP algorithm, which has a computation complexity of O(n^3). Future work aims to reduce this complexity.
1：Experimental Design and Method Selection:
The unmixing problem is modeled as an optimization problem with the objective function based on the secant function. The sequential quadratic programming (SQP) is used to solve the optimization problem.
2：Sample Selection and Data Sources:
Synthetic images with spatial size 50×50 pixels and spectral size l=207, and real hyperspectral dataset (AVIRIS Cuprite dataset) are used.
3：List of Experimental Equipment and Materials:
ENVI software library for synthetic data generation and AVIRIS Cuprite dataset for real data.
4：Experimental Procedures and Operational Workflow:
The proposed method is applied to both synthetic and real datasets, and its performance is compared with state-of-the-art linear unmixing methods using RMSE, AAD, and AE indices.
5：Data Analysis Methods:
The performance is evaluated using root-mean-square error (RMSE), root-mean-square of abundance angle distance (AAD), and abundance error (AE).
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
