Real-Time Hyperspectral Target Detection via Library Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high computational complexity of hyperspectral image processing hinders real-time target detection due to the large volume and complexity of hyperspectral data, requiring efficient data reduction and simplified algorithms.
Innovation Solution
A real-time target detection method that refines a library by extracting effective bands based on a contribution factor, using a pipeline structure and data partitioning to reduce the number of spectral bands and processing complexity, enabling efficient hyperspectral image processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional hyperspectral image processing uses hundreds of bands to detect targets, then detection accuracy is improved, but computational complexity increases making real-time processing impossible
Solution Approach 1:
The patent segments the hyperspectral data by dividing the full spectral range into multiple subsets, processing each subset separately. This segmentation reduces the computational burden of processing all hundreds of bands simultaneously while maintaining detection accuracy through selective band utilization.
Solution Approach 2:
The patent extracts only the most relevant spectral bands for target detection using dimensionality reduction techniques and band selection algorithms. By taking out and processing only the critical bands rather than all bands, the system achieves real-time processing capability while preserving detection accuracy.
2Productivity
If the number of spectral bands is reduced for real-time processing, then processing speed is improved, but detection accuracy may deteriorate
Solution Approach 1:
The patent changes the parameter of spectral band selection dynamically, using adaptive algorithms to identify and process only the most informative bands for each detection scenario. This parameter change enables the system to maintain high detection accuracy with reduced band count, achieving both real-time processing and accurate detection.
3Reliability
If all spectral bands are processed to ensure accurate target detection, then detection reliability is improved, but processing time increases beyond real-time requirements
Solution Approach 1:
The patent performs preliminary actions by pre-processing the hyperspectral data to identify and prioritize the most relevant spectral bands before actual target detection. This preliminary band selection and data preparation reduces the computational workload for the main detection process, enabling real-time processing while maintaining reliable detection results.
Data Source
AI summary
There is provided a method for real-time target detection comprising detecting a preprocessed pixel as a target and/or a background, based on a library, and refining the library by extracting a sample from the target or the background.