Bayer Interpolation Method by Integrating Color-Difference Gradient and Color Correlation

Objectives In order to alleviate the pressure of satellite image data transmission between satellite and ground, the Bayer imaging mode is increasingly being used in remote sensing satellite imaging systems. The information loss occurs because each pixel in the Bayer image only stores information from one color channel. Therefore, a full-color image must be reconstructed by Bayer interpolation for subsequent applications. However, the existing interpolation methods suffer from false colors at image details and overall color tone shift issues.

Methods In response to the above issues, this article proposes a Bayer interpolation method by integrating color-difference gradient and color correlation. First, the proposed method extracts the color difference gradient information between each channel of the original Bayer image, and uses a normalized color difference gradient function to assign corresponding weights to the estimated color difference values in each direction. The interpolation is carried out along the edge direction to protect the detail information at the edge. Then, the color ratio constraints are set according to the correlation between different color channels, and the relative proportion of the three channels in the local sliding window is adjusted to improve the color shift phenomenon in the interpolation results, which makes the reconstructed full-color image have real color tones and complete details.

Results The experimental comparison is conducted on simulated images and real Bayer images obtained from Luojia3-01 satellite, and the proposed method has significantly improved the false color problem at the edges of ground objects. Compared the comparison methods, the proposed method has the highest peak signal-to-noise ratio and structural similarity index, and the lowest mean square error and natural image quality evaluation score.

Conclusions The proposed method can effectively improve the problem of missing details at the edges of ground features and enhance the quality of reconstruction. It has been successfully applied to the data processing system of Luojia3-01 satellite to interpolate Bayer images in orbit and on the ground, providing effective guarantees for the various applications of Luojia3-01 satellite data.