Resolves the SMILE defect in hyperspectral satellite imaging


Our inspiration for SMILE stemmed from the need to address the SMILE (Spectral Miscalibration Identification and Correction) defect prevalent in hyperspectral satellite imaging. The challenge of rectifying this defect, which impacts data accuracy and integrity, motivated our team to develop a solution that enhances spectral calibration.

What it does🔎

SMILE is a novel algorithm designed to rectify the spectral miscalibration issue commonly encountered in hyperspectral satellite imaging. By leveraging a multi-step process, the algorithm corrects the discrepancies in spectral data, enhancing the accuracy and reliability of hyperspectral images.

How we built it🔨

We constructed SMILE utilizing Python and key scientific computing libraries. The algorithm is structured around several crucial steps:

  • Generating Column Averaged Spectra: Calculation of average spectra within each column of the dataset.
  • Spectral Response Function Generation: Creating SRFs for the spectral data.
  • Spectral Angle Calculation: Measuring the spectral angle between test and reference spectra.
  • Correction of Spectral Data: Applying corrective measures to the datacube to rectify spectral miscalibration.

Challenges we ran into⚠️

During the development of SMILE, challenges surfaced in data manipulation, particularly in spectral angle calculation and spline interpolation. Additionally, optimizing computational efficiency in the correction process posed a notable challenge.

Accomplishments that we’re proud of🥇

We successfully implemented SMILE to correct spectral miscalibration, achieving enhanced accuracy and precision in hyperspectral satellite imaging. The algorithm efficiently rectifies spectral discrepancies, laying the groundwork for improved data integrity.

What we learned🧠

The development of SMILE provided valuable insights into spectral analysis and correction methodologies. Our team gained proficiency in utilizing scientific computing tools and techniques, furthering our understanding of hyperspectral imaging correction.

What’s next for SMILE💭

Moving forward, we aim to enhance SMILE by implementing advanced algorithms for spectral correction and exploring opportunities for parallel computing to optimize computational speed.


  • Shivesh Prakash
  • Rosinda Liang
  • Amreen Imrit
  • Shuhan Zheng
  • Andy Gong
  • Ian Vyse
  • Rise Adhikari
  • Julia Ye
  • Rediet Yohannes

Check it out here: GitHub Repo