Mutual information registration of multi-spectral and multi-resolution images of DigitalGlobe's WorldView-3 imaging satellite

Miecznik, Grzegorz; Shafer, Jeff; Baugh, W.; Bader, Brett W.; Karspeck, M.; Pacifici, Fabio

doi:10.1117/12.2262598

Cited by 1 publication

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Foreign studies on the consistency accuracy of satellite multispectral image products have proposed targeted methods. Hayder Dibs et al employed control point first-and second-order polynomial models to enhance the registration accuracy of Razak-SAT satellite images [1]; Grzegorz Miecznik et al achieved high-precision registration between bands in WorldView-3 satellite images using mutual information [2]; Mark A. Gofortl applied derivative algorithms to achieve sub-pixel accuracy in the alignment of four bands in Quickbird and IKONOS satellite images [3]; Issam Boukerch et al achieved sub-pixel registration accuracy using a non-central cross-correlation algorithm for Alsat-2 (A and B) satellites, significantly improving the quality of multispectral images [4]. In China, Wang et al have proposed a satellite multispectral automatic registration method based on geospatial positioning consistency, which enables the automatic registration of multispectral images at the sub-pixel level without requiring image matching [5].…”

Section: Introductionmentioning

confidence: 99%

Joint Panchromatic and Multispectral Geometric Calibration Method for the DS-1 Satellite

Jiang,

Zhang,

Liu

et al. 2024

Remote Sensing

View full text Add to dashboard Cite

The DS-1 satellite was launched successfully on 3 June 2021 from the Taiyuan Satellite Launch Center. The satellite is equipped with a 1 m panchromatic and a 4 m multispectral sensor, providing high-resolution and wide-field optical remote sensing imaging capabilities. For satellites equipped with panchromatic and multispectral sensors, conventional geometric processing methods in the past involved separate calibration for the panchromatic sensor and the multispectral sensor. This method produced distinct internal and external calibration parameters in the respective bands, and also resulted in nonlinear geometric misalignments between the panchromatic and multispectral images due to satellite chattering and other factors. To better capitalize on the high spatial resolution of panchromatic imagery and the superior spectral resolution of multispectral imagery, it is necessary to perform registration on the calibrated panchromatic and multispectral images. When registering separately calibrated panchromatic and multispectral images, poor consistency between panchromatic and multispectral images leads to a small number of corresponding points, resulting in poor accuracy and registration effects. To address this issue, we propose a joint panchromatic and multispectral calibration method to register the panchromatic and multispectral images. Before geometric calibration, it is necessary to perform corresponding points matching. When matching, the small interval between the panchromatic and multispectral Charge-Coupled Devices (CCDs) results in a small intersection angle of the corresponding points between the panchromatic and multispectral images. As a result of this, the consistency between the spectral bands significantly improves, and the corresponding points match to have a more uniform distribution and a wider coverage. The technique enhances the consistent registration accuracy of both the panchromatic and multispectral bands. Experiments demonstrate that the joint calibration method yields a registration accuracy of panchromatic and multispectral bands exceeding 0.3 pixels.

show abstract