Multi-band Hyperspectral Imaging System for Machine Vision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hyperspectral imaging systems are bulky, costly, and unsuitable for consumer applications due to their complex optical components and high spectral resolution, which results in insufficient spatial resolution and high production costs, making them impractical for machine vision applications like UAVs and robotics.
Innovation Solution
A novel hyperspectral imaging system that uses all camera sensor pixels to capture images in different spectral bands with bandpass filters positioned in front of the image sensors, offering higher spatial and spectral resolution, and is designed to be lightweight, cost-effective, and suitable for mass production.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional hyperspectral imaging systems use high spectral resolution (around 3 nm), then spectral measurement precision is improved, but spatial resolution becomes insufficient and system complexity increases
Solution Approach 1:
The imaging system divides the spectral response range into multiple discrete bands using separate bandpass filters for each camera sensor. Each filter is optimized for its specific spectral band (e.g., 400-500nm, 500-600nm, 600-700nm), allowing each sensor to capture its designated spectral range with high precision while maintaining adequate spatial resolution through parallel processing of multiple bands simultaneously
Solution Approach 2:
The system transitions from sequential scanning methods (point-scan or line-scan) to a parallel multi-band approach where multiple spectral bands are captured simultaneously across the entire field of view. This dimensional change from time-based scanning to space-based parallel measurement resolves the contradiction by capturing both high spectral and spatial resolution data at the same time
2Measurement precision
If conventional hyperspectral systems use multiple optical spectroscopic components, then spectral resolution is improved, but device complexity and cost increase
Solution Approach 1:
Instead of using complex optical spectroscopic components to achieve spectral resolution, the system uses multiple copies of simple bandpass filters, each optimized for a specific spectral band. These filters are positioned in front of corresponding camera sensors, replacing complicated optical path systems with simpler, more cost-effective filter-based spectral separation
Solution Approach 2:
Each camera sensor in the system serves multiple functions: it captures spatial information across the entire field of view while simultaneously providing spectral information for its designated band. This multi-functionality eliminates the need for separate optical spectroscopic components, reducing overall system complexity while maintaining spectral resolution
3Measurement precision
If point-scan or line-scan imaging systems scan the entire spectral response range, then spectral resolution is improved, but imaging time becomes excessively long
Solution Approach 1:
The system maintains continuous imaging by capturing all spectral bands simultaneously across the entire field of view in a single operation. Unlike sequential scanning methods that pause imaging to scan through spectral ranges, this parallel approach keeps the camera sensors continuously capturing data without interruption, thereby maintaining high productivity while achieving high spectral resolution
4Device complexity
If conventional hyperspectral systems use single optical lens systems, then device complexity is reduced, but spatial resolution becomes insufficient
Solution Approach 1:
The system divides the imaging function across multiple camera sensors, each with its own bandpass filter. This segmentation allows each sensor to independently capture high-resolution spatial data for its designated spectral band, achieving overall high spatial resolution without requiring a single complex optical lens system
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
The system provides improved spatial and spectral resolution, enabling real-time, cost-effective hyperspectral imaging suitable for various machine vision applications, including UAVs, robotics, and autonomous driving, while being more affordable and compact than traditional systems.
Implementation Method 1
Each bandpass filter has a passing band that allows transmission of incident electromagnetic signals between a first and second wavelength associated with the filter
Data Source
AI summary
A novel imaging system is provided for capturing imaging data. According to the disclosed embodiments, the imaging system comprises one or more cameras configured to capture the same scene within each camera's field of view. The imaging system also includes a plurality of bandpass filters that may be positioned in front of one or more cameras. Each bandpass filter may allow the transmission of incident electromagnetic signals between different pairs of first and second wavelengths. Accordingly, when the bandpass filters are positioned in front of the one or more cameras, each camera captures a different spectral image, i.e., containing only certain frequency components of the same scene being imaged in the cameras. The bandpass filters may be selectively aligned with one or more cameras by rotating and/or translating the filters relative to the cameras' positions in the imaging system.