Adaptive Bad Pixel Correction Method for Interference-Modulated Images Based on Weighted Least Squares Support Vector...

Area of Science:

• Remote Sensing
• Spectroscopy
• Image Processing

Background:

• Temporally and spatially modulated Fourier transform imaging spectrometers (TSMFTISs) are crucial for remote sensing, enabling target classification and identification through spectral information.
• Defect pixels in planar array charge-coupled devices significantly degrade the accuracy of spectral recovery in TSMFTIS data.
• Preprocessing of bad pixels is essential for reliable data processing in TSMFTIS.

Purpose of the Study:

• To address the impact of defect pixels on TSMFTIS data accuracy.
• To introduce an adaptive defect pixel correction method for TSMFTIS.
• To evaluate the efficiency and robustness of the proposed method compared to traditional approaches.

Main Methods:

• An adaptive defect pixel correction algorithm based on weighted least squares support vector machine (WLS-SVM) was developed.
• The principle of TSMFTIS was analyzed to understand the nature of bad pixels and limitations of existing methods.
• Simulations were conducted to compare the proposed WLS-SVM method with conventional techniques for bad pixel correction.

Main Results:

• The proposed adaptive method demonstrates superior efficiency and robustness in correcting defect pixels for TSMFTIS.
• The WLS-SVM approach effectively mitigates the negative impact of absent pixels on target spectral recovery.
• The study validates the practical application of the developed defect pixel correction method.

Conclusions:

• The developed adaptive defect pixel correction method is highly effective for TSMFTIS in remote sensing applications.
• The WLS-SVM algorithm offers a significant improvement over traditional methods for spectral data preprocessing.
• Accurate spectral recovery is achievable with robust defect pixel correction in TSMFTIS.