14:00 〜 14:15
[U03-02] A color-matching method and application for Chinese HY-1 imagery in Antarctica
キーワード：HY-1C/D, Antarctic, color-matching, image mosaic
Mapping the Antarctic using satellite imagery is a crucial and foundational task in polar environmental monitoring. It enables the study of large-scale environmental changes, such as the monitoring of glacial mass balance, ice shelf calving, and Antarctic ecology, aiding in the understanding of the impact of polar changes on global climate. The Coastal Zone Imager (CZI) aboard the HY-1C/1D satellites, part of the Chinese ocean color satellite series, is an optical remote sensing payload. It features blue, green, red, and near-infrared channels with a spatial resolution of 50 meters and a swath width of 900 kilometers. The enhanced imaging capabilities of the CZI offer an efficient observational tool for the Antarctic region.
Image mosaicking is a technique for seamlessly fusing multiple images with overlapping areas into a single composite image. Ensuring color consistency between different images is a crucial step in the image mosaicking process, aimed at eliminating color discrepancies and achieving uniformity across the mosaic. In this paper, we explore the illumination homogeneity correction of individual images and the color registration of multiple images using data from Antarctic CZI images. We propose a gray-level segmentation color-mapping method to address the issue of image overstretch during color-matching in the Antarctic region, which is typically a problem with traditional histogram matching methods. Statistical results indicate that our proposed method effectively eliminates color bias between the CZI Antarctic experimental images.
Ultimately, we obtained a dataset comprising 251 Level 1C (L1C) images captured by the HY-1C/D CZI satellite over the Antarctic region, covering the period from October 2021 to November 2024. After undergoing rigorous cloud removal procedures and the application of the color registration technique described in this study, we successfully produced a seamless mosaic image covering the entire Antarctic continent. The results indicate that the methodology employed in this research demonstrates outstanding applicability for the color registration of HY-1 satellite imagery in the Antarctic region, generating visually appealing and satisfactory outcomes.
Image mosaicking is a technique for seamlessly fusing multiple images with overlapping areas into a single composite image. Ensuring color consistency between different images is a crucial step in the image mosaicking process, aimed at eliminating color discrepancies and achieving uniformity across the mosaic. In this paper, we explore the illumination homogeneity correction of individual images and the color registration of multiple images using data from Antarctic CZI images. We propose a gray-level segmentation color-mapping method to address the issue of image overstretch during color-matching in the Antarctic region, which is typically a problem with traditional histogram matching methods. Statistical results indicate that our proposed method effectively eliminates color bias between the CZI Antarctic experimental images.
Ultimately, we obtained a dataset comprising 251 Level 1C (L1C) images captured by the HY-1C/D CZI satellite over the Antarctic region, covering the period from October 2021 to November 2024. After undergoing rigorous cloud removal procedures and the application of the color registration technique described in this study, we successfully produced a seamless mosaic image covering the entire Antarctic continent. The results indicate that the methodology employed in this research demonstrates outstanding applicability for the color registration of HY-1 satellite imagery in the Antarctic region, generating visually appealing and satisfactory outcomes.