Spectral Sorting of Glass Using Monochrome Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current bulk material sorting systems, particularly those using color cameras, face challenges with label interference, low spatial resolution, and high over-sorting rates, especially with dark and thick colored glass, leading to inefficient and economically unsatisfactory sorting outcomes.
Innovation Solution
A spectral-based sorting method utilizing a monochrome sensor with illumination at predeterminable wavelengths in the UV, VIS, and NIR ranges, combined with a partial dark field lighting system, allows for reliable detection and sorting of transparent materials, including glass with labels and curved objects, by evaluating transmission and absorption ratios across multiple spectral ranges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If color camera technology is used for sorting, then color sorting capability is improved, but labels on materials interfere with detection and cause high over-sorting rates
Solution Approach 1:
The spectrum is segmented into multiple discrete wavelength ranges (UV, VIS, NIR) with specific wavelength points selected for measurement. This segmentation allows the system to measure transmission at multiple spectral points, enabling differentiation between labels and colored glass by analyzing the distinct transmission patterns at each wavelength segment.
Solution Approach 2:
A multi-wavelength light source system acts as an intermediary between the material and sensor. By illuminating the material with multiple discrete wavelengths and measuring transmission at each, the system creates a spectral fingerprint that mediates the detection process, allowing differentiation between labels and colored glass through their distinct absorption characteristics at different wavelengths.
2Quantity of substance
If transmitted light sorting is used for transparent materials, then glass and transparent plastic products can be sorted, but dark and thick colored glass have low transmission and are sorted into non-transparent materials
Solution Approach 1:
The system transitions from measuring only visible light transmission to measuring transmission across multiple dimensional spectral ranges (UV, VIS, NIR). By adding the NIR dimension, dark and thick colored glass that appear opaque in the visible range can be detected through their transmission characteristics in the NIR range, preventing their misclassification as non-transparent materials.
Solution Approach 2:
The measurement parameters are changed from single-wavelength or broad-spectrum visible light measurement to multi-wavelength measurement across UV, VIS, and NIR ranges. This parameter change allows the system to detect transmission variations at different energy levels, enabling accurate identification of dark colored glass by its characteristic absorption patterns across the spectral range.
3Reliability
If NIR-based devices are used for sorting non-transparent impurities, then recognition of ceramics, stones and porcelain is improved, but color sorting capability is lost and undesired cycles occur
Solution Approach 1:
The multi-wavelength illumination and measurement system is designed to perform multiple functions: it can detect transparent materials, identify colored glass, recognize non-transparent impurities, and differentiate between them all within a single sorting pass. The system's ability to measure across UV, VIS, and NIR ranges provides universal detection capability that eliminates the need for separate NIR-only systems and prevents material cycling.
4Measurement precision
If high spatial resolution and high line rate are achieved with color cameras, then sorting precision is improved, but device complexity and cost increase
Solution Approach 1:
The system extracts only the essential spectral information needed for sorting by measuring at specific discrete wavelength points rather than capturing the entire spectrum. This extraction approach reduces data processing complexity while maintaining sufficient measurement precision for accurate material identification and sorting.
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 enables accurate and efficient sorting with reduced over-sorting, improving material identification and reducing material waste by distinguishing between different materials and colors, even in complex scenarios with labels and broken edges.
Implementation Method 1
a lighting unit (2) for illumination with predeterminable wavelengths λ
Implementation Method 2
the sensor (1) is designed to record transmission signals
Implementation Method 3
combined with a partial dark field lighting system
Data Source
Figure 1~2
Figure 3~4
AI summary
The method involves moving bulk goods (8, 9) in a material stream between an illumination unit (2) and a sensor (1) i.e. monochrome sensor, which is designed as a line scan camera or an area scan camera. Illumination with a predetermined wave length is implemented in a temporally or locally discrete manner using the illumination unit. Illumination is produced as a partial dark field illumination. Transmission signals are received by the sensor. The transmission signals are evaluated and bulk goods are sorted based on the evaluation. The illumination unit comprises light elements (3) i.e. high power LEDs, which are arranged outside a visible area (4) of the illumination unit. An independent claim is also included for a device for spectral based sorting of transparent and semitransparent bulk goods.