Textile Classification Using Proximity-Triggered NIR/SWIR Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-invasive textile substrate classification methods lack precise procedures for reflectance and transmittance spectrum processing and feature extraction, leading to potential damage from improper laundry process parameters and inadequate material verification in home appliances and textile manufacturing.
Innovation Solution
A proximity triggered textile substrate classification apparatus using NIR/SWIR spectroscopy with integrated proximity sensors, aspheric lenses, and a central processor executing a fabric classification algorithm that includes spectral reflectance measurement, feature extraction via smoothing, normalization, and decorrelation using DCT, followed by machine learning for accurate classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If non-invasive textile substrate classification is performed using infrared spectroscopy, then the garment integrity is preserved, but the measurement precision and classification accuracy are insufficient due to lack of precise spectrum processing procedures
Solution Approach 1:
The patent applies preliminary action by performing proximity detection before spectral measurement to ensure optimal positioning. The system detects when the textile substrate is at the correct distance from the sensor array using proximity sensors, triggering the measurement only when conditions are optimal. This preliminary positioning step ensures that subsequent spectral measurements are performed with maximum precision while maintaining non-invasive conditions.
Solution Approach 2:
The patent replaces mechanical contact-based classification methods with optical field-based infrared spectroscopy. Instead of physically touching or cutting the garment to analyze its composition, the system uses infrared light interaction with the textile substrate. This substitution of mechanical analysis with optical field analysis preserves garment integrity while enabling accurate compositional classification through spectral analysis.
2Productivity
If spectral measurements are performed without proximity triggering, then the measurement speed is faster, but the signal-to-noise ratio deteriorates due to inconsistent positioning
Solution Approach 1:
The patent implements feedback through proximity sensors that continuously monitor the distance between the sensor array and the textile substrate. The system uses this proximity feedback to trigger measurements only when optimal positioning is detected, ensuring high signal-to-noise ratio. The feedback loop automatically adjusts measurement timing based on real-time positioning information, maintaining both speed and precision.
Solution Approach 2:
The system employs periodic action by using the proximity trigger to initiate spectral measurements at optimal intervals. Instead of continuous measurement, the system periodically activates the spectrometer array only when proximity conditions are met, which optimizes the balance between measurement speed and signal quality. This periodic triggering based on proximity events maintains high productivity while ensuring each measurement has adequate signal-to-noise ratio.
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
Ensures precise and non-destructive textile substrate identification, improving laundry process efficiency and material verification by enhancing signal-to-noise ratio and reducing classification complexity, thereby minimizing damage and ensuring accurate material recognition.
Implementation Method 1
The apparatus also contains a proximity sensor 4, which is one of the crucial components of the apparatus according to this invention. The purpose of the proximity sensor is to detect the presence of the measured material 1 (textile substrate) near the predefined distance from the VIS and NIR/SWIR aperture 3.
Implementation Method 2
Infrared spectroscopy (especially near (NIR) and shortwave (SWIR) infrared spectroscopy) was found to be the most suitable general method for non-invasive textile substrate classification. Namely, different textile substrates were found to have rather specific IR reflectance characteristics. Therefore, the textile substrate classification can be performed by illuminating the analysed textile substrate with infrared light source and afterwards carefully analysing the IR light, which diffusely reflects from the analysed textile substrate.
Implementation Method 3
the IR light, which diffusely reflects from the analysed textile substrate
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to apparatus and procedure for non-invasive textile substrate classification. The apparatus includes the proximity sensor (4) to trigger the reflective NIR/SWIR spectral response measurement, feature extraction, and textile substrate machine classification. Instead of the direct spectrum output, the cepstral features are first derived from the post-processed NIR/SWIR spectral response and only then used in the machine classification. The optical front-end contains a setup consisting of the two aspheric lenses (12,13) with different numerical apertures and the physical aperture positioned between the two, such that the distance to each lens (12,13) matches the focal length of the respective lens. In one embodiment of the invention, the textile substrate classification procedure is described in hardware description layer (HDL) thus running inside the Field Programmable Gate Array (FPGA) integrated circuit.