Hyperspectral Imaging With Holographic Optics for Material Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for material classification at the molecular level are expensive, require expert knowledge, and are unable to optimally use the area of image sensors, limiting their ability to classify specific materials or map each spectra to an area of the image sensor.
Innovation Solution
The use of holographic optical elements that filter and reflect specific wavelengths of light, combined with machine learning, to enhance the classification process by mapping each spectrum to an area of the image sensor, allowing for the authentication and classification of molecular compositions of products.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional spectroscopy systems are used for material classification, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent segments the continuous spectral region into multiple discrete spectral bands, with each band captured by a dedicated region on the image sensor. This segmentation allows the system to focus on specific wavelength ranges relevant to different molecular signatures, achieving molecular-level classification precision while using a simpler, more compact sensor array architecture rather than a complex continuous scanning spectrometer
Solution Approach 2:
The patent transitions from traditional one-dimensional spectral scanning to a two-dimensional spatial-spectral mapping approach. By mapping different spectral bands to different spatial regions on the image sensor, the system captures multiple spectral measurements simultaneously in parallel, dramatically reducing system complexity and enabling real-time molecular classification without mechanical scanning components
2Measurement precision
If traditional spectroscopy systems are used for material classification, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent implements machine learning algorithms that automatically analyze the spectral data captured by the image sensor and classify materials based on their molecular signatures. The system self-calibrates and self-optimizes by comparing spectral patterns against reference databases, eliminating the need for users to have expert knowledge of spectroscopy interpretation or manual calibration procedures
Solution Approach 2:
The patent replaces complex mechanical scanning spectrometers with a static, solid-state image sensor array that captures spectral information in parallel. This substitution of mechanical systems with electronic sensors simplifies the device structure, reduces moving parts, and enables easier operation without requiring users to understand or manipulate complex optical scanning mechanisms
3Productivity
If image sensors are used for spectral capture, then productivity is improved, but measurement precision deteriorates due to limited resolution
Solution Approach 1:
The patent segments the spectral range into multiple discrete bands, with each band assigned to a specific region on the image sensor. This segmentation allows the system to capture multiple spectral measurements simultaneously in parallel across different sensor regions, achieving high productivity while maintaining sufficient precision for molecular classification by focusing on specific wavelength ranges rather than attempting to resolve the entire spectrum at high detail
Solution Approach 2:
The patent combines multiple spectral measurements taken simultaneously by different regions of the image sensor to create a comprehensive spectral profile. By merging the parallel spectral data from multiple sensor regions, the system achieves both high productivity through simultaneous capture and adequate precision through the integration of multiple spectral bands for molecular signature identification
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 provides a cost-effective and user-friendly method for authenticating products by classifying molecular compositions, optimizing the use of image sensors and enabling accurate classification of materials such as wine, while being accessible to consumers.
Implementation Method 1
a holographic printer is used to print an optical element, the holographic optical element, an array of pixels composing the optical element are specified to reflect a specific set of line spectra useful for classifying and authenticating a set of materials or products
Implementation Method 2
an array of pixels composing the optical element are specified to reflect a specific set of line spectra
Data Source
AI summary
In some implementations, a method comprises: determining a set of characteristics of light spectra reflected or transmitted by a set of materials when the set of materials is illuminated by a plurality of light wavelengths; constructing one or more classifiers configured to classify each material of the set of materials based on the set of characteristics of the light spectra; using the classifiers, mapping each of the light spectra onto an area of an image sensor; wherein one or more optical elements filter and focus the light spectra onto one or more elements of a detector array; wherein each optical element focuses onto an area of the detector array; wherein the mapping is a 1:1 mapping; wherein each optical element uses a Bragg reflection condition to filter the light spectra and to focus the light spectra onto an element of the detector array.