Ocean Surface Target Visibility via Gravity Wave Propagation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current deglinting algorithms fail to effectively remove wave clutter due to time delays between image pairs, leading to misalignment and incomplete removal of wave patterns, especially in satellite imaging where time differences between spectral bands can be significant.
Innovation Solution
The method involves obtaining a pair of images with pixel intensities proportional to wave height, aligning and orthorectifying them, and using Fourier transforms to propagate one image to the time of the other, with a gravity wave propagation function accounting for depth, surface current, and wave direction, allowing for accurate subtraction of wave patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional deglinting algorithms are used with time-delayed image pairs, then the processing is simpler, but the wave clutter removal is incomplete due to misalignment
Solution Approach 1:
The patent applies wave propagation modeling in advance to predict how wave patterns evolve between the two image acquisition times. By pre-calculating the propagation effect and applying it to the first image before subtraction, the algorithm compensates for temporal misalignment without requiring complex real-time adjustments during the deglinting process itself.
Solution Approach 2:
The patent introduces temporal parameters (time delay between images, wave propagation speed, wave period) into the deglinting process. By varying these parameters and selecting optimal values that match the actual wave conditions, the algorithm achieves accurate wave pattern alignment despite the time delay, thereby improving clutter removal precision.
2Adaptability or versatility
If image pairs with significant time delay are used, then satellite imaging can capture different spectral bands, but wave patterns become misaligned reducing deglinting effectiveness
Solution Approach 1:
The patent performs wave propagation compensation before the image subtraction step. By predicting the wave pattern evolution over the known time delay period and applying this prediction to align the first image with the second, the method ensures accurate wave pattern matching despite significant time delays between spectral band acquisitions.
Solution Approach 2:
The patent creates a propagated copy of the first image's wave pattern that reflects how the waves would appear at the later time of the second image acquisition. This propagated copy is then used in place of the actual later wave pattern for subtraction, effectively creating a virtual alignment that accounts for temporal changes.
3Measurement precision
If wave propagation effects are accounted for, then target visibility is improved, but processing complexity increases
Solution Approach 1:
The patent replaces complex mechanical or iterative alignment methods with a mathematical wave propagation model. Instead of trying various alignment transformations or performing complex optimization to match wave patterns, the method uses analytical solutions to wave propagation equations to directly calculate the expected wave pattern evolution, simplifying the processing while improving accuracy.
Solution Approach 2:
The patent transforms the complex wave alignment problem into a parameter estimation problem. By identifying key wave parameters (period, direction, speed) and using these to drive the propagation model, the method reduces the complexity of handling full wave fields while maintaining accurate alignment for target visibility enhancement.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach results in improved visibility of targets on or below the water surface by accurately accounting for time shifts and varying conditions, such as depth and currents, enhancing the deglinting process and reducing noise, thereby improving target recognition and imaging quality.
Implementation Method 1
The propagation uses a gravity wave propagation function, which, in general, can be a function of depth, surface current, and wave direction
Implementation Method 2
The images are Fourier transformed, and then one image is propagated to the time of the other image
Data Source
AI summary
A pair of images is obtained having pixel intensities substantially proportional to wave height for a majority of pixels. The images are aligned, orthorectified, and adjusted for relative reflectance in their respective spectral bands. The images are Fourier transformed, and then one image is propagated to the time of the other image. A deglinted image is calculated by taking the difference in wavenumber space, then inverse Fourier transforming. Alternatively, an equivalent formula can be obtained using convolutions. The propagation uses a gravity wave propagation function, which, in general, can be a function of depth, surface current, and wave direction which can be obtained from external data, assumed constant, or calculated from wave images depending on the particular circumstances.