Hyperspectral Imaging System Using Compressive Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current hyperspectral imaging systems are expensive and inefficient, requiring multiple sensors and struggling with consistent calibration, especially when imaging dynamic scenes or those affected by ambient light changes.
Innovation Solution
The system employs input optics, spatial and spectral modulators, and a single pixel sensor with electronic processing to generate hyperspectral images through spatial and spectral sampling, using techniques like compressive sensing and singular value decomposition to reduce the number of required samples and improve imaging speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If spatial scanning using 1-D or single pixel sensor is used, then imaging time is reduced, but measurement precision and reliability deteriorate due to limited sampling
Solution Approach 1:
The patent applies preliminary action by performing compressive sensing measurements that capture essential spectral information in advance, allowing accurate reconstruction from fewer samples. The system pre-processes the spectral data through coded aperture or spatial light modulator patterns before detection, enabling high-speed imaging without sacrificing spectral fidelity.
Solution Approach 2:
The patent changes the sampling parameter regime by using compressive sensing to acquire images at sub-Nyquist rates. By transforming the traditional uniform sampling approach into compressed random sampling with structured masks, the system achieves both high speed and accurate spectral reconstruction from reduced measurements.
2Measurement precision
If multiple sensors are used to improve measurement precision, then device complexity and cost increase
Solution Approach 1:
The patent merges multiple sensing functions into a single pixel sensor by using spatial light modulators or coded aperture masks that encode spectral information spatially before detection. This consolidation maintains spectral reconstruction accuracy while eliminating the need for multiple physical sensors, thereby reducing system complexity and cost.
Solution Approach 2:
The patent introduces spatial light modulators or coded aperture masks as intermediary elements between the scene and the single pixel sensor. These intermediaries perform spectral encoding and spatial modulation, enabling a single sensor to capture information that would traditionally require multiple sensors, thus simplifying the overall system architecture.
3Loss of information
If traditional hyperspectral imaging is used, then spectral information is complete, but imaging time becomes unacceptably long for dynamic scenes
Solution Approach 1:
The patent applies partial action by acquiring only the essential spectral information needed for accurate reconstruction rather than complete spectral sampling. Compressive sensing allows the system to capture a subset of spectral measurements that, when properly processed, retain sufficient information to reconstruct the full spectrum, thereby reducing acquisition time for dynamic scenes.
Solution Approach 2:
The system performs preliminary compressive measurements that capture the most significant spectral features before the scene changes. By using structured sampling patterns and prior knowledge of spectral sparsity, the system prepares and acquires critical information in advance, enabling accurate reconstruction even from limited, time-compressed measurements.
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 enables rapid hyperspectral imaging with reduced costs and sample requirements, achieving high accuracy and efficiency by processing signals to reconstruct images with fewer measurements, thus overcoming the limitations of traditional systems.
Implementation Method 1
a spatial modulator that spatially samples radiation received from the input optics to generate spatially sampled radiation
Implementation Method 2
a spectral modulator that spectrally samples the spatially sampled radiation received from the spatial modulator to generate spectrally sampled radiation
Implementation Method 3
a sensor that senses spectrally sampled radiation received from the spectral modulator and generates a corresponding output signal
Implementation Method 4
compressive sampling takes advantage of a signal's sparseness or compressibility in some domain to allow the entire signal to be recreated from an even smaller number of measurements
Implementation Method 5
Signal processing is often used to reconstruct a signal from a series of sampling measurements
Data Source
AI summary
Apparatus for hyperspectral imaging, the apparatus including input optics that receive radiation reflected or radiated from a scene, a spatial modulator that spatially samples radiation received from the input optics to generate spatially sampled radiation, a spectral modulator that spectrally samples the spatially sampled radiation received from the spatial modulator to generate spectrally sampled radiation, a sensor that senses spectrally sampled radiation received from the spectral modulator and generates a corresponding output signal and at least one electronic processing device that controls the spatial and spectral modulators to cause spatial and spectral sampling to be performed, receives output signals and processes the output signals in accordance with performed spatial and spectral sampling to generate a hyperspectral image.