Multiplexed Compressive Imaging for Low-Data Optical Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital image and video acquisition methods require large amounts of raw data, which are expensive to obtain and computationally demanding to compress, especially at wavelengths where CMOS or CCD sensing technology is limited, and existing compressive sensing techniques are not universally applicable across the electromagnetic spectrum.
Innovation Solution
An imaging system utilizing a multilayered modulator to directly acquire a compressed digital representation of a signal by modulating an incident light field with pseudorandom patterns, optically computing inner products, and employing algorithms like Greedy reconstruction or Basis Pursuit to recover the signal, applicable across various electromagnetic spectrum ranges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional digital image or video acquisition methods are used, then complete raw data is obtained, but the data volume becomes large requiring expensive sensing technology and computationally demanding compression
Solution Approach 1:
The patent applies preliminary action by performing compression during the data acquisition process itself rather than after acquisition. The compressive sensing framework encodes the scene directly into a compressed representation during measurement, eliminating the need for subsequent compression steps and reducing overall data volume while maintaining reconstruction quality
Solution Approach 2:
The patent extracts only the essential information from the scene through compressive measurements. By using random projection matrices and exploiting signal sparsity, the system extracts the minimum necessary measurements to reconstruct the image, discarding redundant information at the acquisition stage rather than storing and then compressing complete raw data
2Loss of information
If large amounts of raw image or video data are acquired, then complete information is captured, but the acquisition cost increases particularly at wavelengths where CMOS or CCD sensing technology is limited
Solution Approach 1:
The system performs information extraction and compression during the measurement process itself. By using compressive sensing with random projections, the essential scene information is captured in a compressed form directly during acquisition, eliminating the need to first acquire large volumes of raw data that would then require expensive specialized sensors for certain wavelengths
Solution Approach 2:
The patent changes the measurement parameters from conventional pixel-by-pixel sampling to compressive measurements using random projection matrices. This parameter change allows the system to acquire fewer measurements that still contain sufficient information for high-quality reconstruction, reducing the need for expensive high-resolution sensors particularly at non-standard wavelengths
3Quantity of substance
If raw data compression is performed, then data volume is reduced, but computational demands increase particularly in the case of video
Solution Approach 1:
The patent performs compression during the acquisition phase rather than as a separate post-processing step. By encoding the scene directly into a compressed representation through compressive measurements, the system reduces data volume at the source, eliminating the need for computationally intensive compression algorithms to be applied later to large video datasets
Solution Approach 2:
The patent replaces traditional mechanical compression systems with an optical-computational approach. The compressive sensing framework uses random projection matrices and sparsity-exploiting algorithms to perform compression during measurement, substituting the need for separate mechanical or software-based compression stages that would demand high computational power for video processing
4Productivity
If existing compressive sensing techniques are used, then fewer measurements are required, but applicability across the electromagnetic spectrum is limited
Solution Approach 1:
The patent achieves universality by developing a compressive sensing framework that can be applied across different wavelengths and sensing modalities. The random projection-based measurement approach and sparsity-exploiting reconstruction algorithms are wavelength-agnostic, allowing the same fundamental technique to be used for visible, infrared, and other electromagnetic spectrum ranges with appropriate sensor selection
Data Source
AI summary
Compressive imaging apparatus employing multiple modulators in various optical schemes to generate the modulation patterns before the signal is recorded at a detector. The compressive imaging apparatus is equally valid when applying compressive imaging to structured light embodiments where the placement is shifted from the acquisition path between the subject and the detector into the illumination path between the source and the subject to be imaged.