VIIRS Image Processing Bow-Tie Artifact Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current VIIRS image processing methods, such as those using the Python SatPy package, are inefficient in removing 'bow-tie' artifacts and spatial discontinuities, leading to slow processing speeds and reduced accuracy in machine learning applications due to inadequate interpolation and geospatial metadata requirements.
Innovation Solution
The proposed solution involves pixel-wise interpolation using matrix multiplication, which corrects spatial discontinuities without requiring geospatial metadata, allowing for faster processing of VIIRS Sensor Data Records (SDRs) and producing high-quality images suitable for machine learning integration by estimating relative ground distance based on sensor sampling characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current SatPy package processing is used, then geospatial metadata requirements are met, but processing time is excessive (3 minutes)
Solution Approach 1:
The patent extracts and removes the geospatial metadata requirement from the processing pipeline, demonstrating that high-quality interpolation and bow-tie artifact removal can be achieved using only sensor sampling characteristics. This eliminates the need for separate geospatial files while maintaining output quality.
Solution Approach 2:
The patent segments the processing into distinct stages: (1) selecting pixels based on relative ground distance, (2) resampling to uniform columns, (3) interpolating missing values, and (4) removing bow-tie artifacts. This modular approach enables optimized processing without geospatial metadata.
2Manufacturing precision
If re-interpolation is performed on multiple SDRs grouped into granules, then ground-track Mercator projection is achieved, but processing complexity and time increase significantly
Solution Approach 1:
The patent performs preliminary selection of pixels based on relative ground distance calculations using only sensor sampling characteristics. This pre-processing step establishes the foundation for subsequent interpolation without requiring complex geospatial projections or multiple SDR groupings.
Solution Approach 2:
The patent changes the approach from geospatial coordinate-based processing to sensor-sampling-characteristic-based processing. By using relative ground distance derived from sensor geometry rather than geospatial metadata, the method simplifies the processing pipeline while maintaining accuracy.
3Reliability
If SatPy re-interpolation is applied, then some spatial discontinuities are reduced, but interpolation quality and speed are insufficient
Solution Approach 1:
The patent replaces the mechanical re-interpolation process of SatPy with a more efficient algorithm that selects pixels based on relative ground distance and uses targeted interpolation only where needed. This substitution achieves better spatial continuity with significantly improved processing speed.
Solution Approach 2:
The patent applies interpolation locally only to columns where spatial discontinuities exist, rather than uniformly processing all pixels. By identifying and treating only the affected regions based on sensor sampling patterns, the method achieves high spatial continuity while minimizing processing time.
4Productivity
If bow-tie artifacts are not removed, then processing speed is maintained, but machine learning application accuracy is reduced
Solution Approach 1:
The patent removes bow-tie artifacts through a continuous process that operates on the interpolated data, ensuring that the entire processing pipeline (selection, resampling, interpolation, and artifact removal) works together seamlessly. This maintains processing speed while eliminating artifacts that would otherwise degrade machine learning accuracy.
Data Source
AI summary
Systems and methods for VIIRS image processing. The method can include receiving image data of immediately adjacent VIIRS image scans including a first image scan and a second image scan. The first image scan and the second image scan provide a partially overlapping view of a geographic area. The method can further involve resampling columns of pixels of the first image scan and the second image scan. The resampling can include selecting, in the first image scan and the second image scan, a subset of pixel values in each column that correspond to a specified geographic distance. The method can further involve upsampling the selected pixels to an equal number of pixels in each column resulting in upsampled pixel values and interpolating the upsampled pixel values to produce modified first and second image scans.