Optical sorting systems make thousands of decisions per minute based on visual or spectral information. While sensors and software often receive the most attention, their performance ultimately depends on the quality of the optical assembly supplying the image data.
How optical sorting systems operate
At a basic level, every optical sorting system follows the same chain:
Imaging/sensing stage → optical assembly → processing → ejection.
Depending on the application, light may be reflected, transmitted, or absorbed by the object before being captured by an imaging or sensing element.
Before reaching the sensor, the light passes through an optical assembly comprising lenses, filters, and, in some cases, beamsplitters or waveplates, all working together to deliver a clean, accurately mapped image or spectral signature. That data feeds a processing system, which makes the sort decision, and finally an ejection mechanism (an air jet, a diverter, a robotic arm) acts on that decision in real time.
The performance of each stage depends on the quality of the data generated by the previous one. A processing algorithm is sophisticated, but it's working entirely from what the optical assembly hands it. The performance of the processing system is limited by the quality of the optical data it receives.
Key Optical Requirements
Lens speed and light-gathering capability: As line speeds increase, exposure time decreases. Optical systems must therefore maximize light collection and transmission efficiency to maintain image quality.
Field of view and depth of field: Variations in product size, geometry, and position place demands on both field of view and depth of field.
Spectral response: Depending on the application, sorting may rely on visible, near-infrared (NIR), or hyperspectral imaging. The optical assembly must be designed for the required wavelength range, as both transmission efficiency and aberration correction vary across the spectrum.
Distortion control: Accurate sorting depends on accurate spatial mapping. Uncorrected lens distortion can introduce positional errors, particularly at the edge of the field of view, which may affect sorting accuracy.
Optical requirements by industry
Food sorting: High-throughput color and defect detection is often the primary challenge. Systems must identify discoloration, blemishes, and foreign material on fast-moving product streams while maintaining tight acceptance tolerances.
Recycling: Applications often rely on spectral analysis to distinguish between different material types, such as polymer grades, metals, and composite materials.
Minerals and pharmaceuticals: These applications often require the detection of subtle compositional or structural differences, demanding both high spatial resolution and precise spectral discrimination at production volumes where consistency and repeatability are critical.
While the requirements vary by application, sorting performance ultimately depends on the quality and consistency of the optical data collected.
In high-speed sorting systems, these optical requirements must be balanced simultaneously. Increasing field of view can affect resolution, while improving depth of field often requires additional illumination. The optical assembly therefore needs to be designed around the application's throughput, detection requirements, and operating environment rather than a single performance metric.
The Importance of optical assembly design
Improvements in sorting performance are often sought through sensors and processing algorithms. In many systems, optical performance ultimately determines the achievable level of accuracy. Even advanced cameras and processing algorithms cannot fully compensate for chromatic aberration, image distortion, or insufficient light collection at high operating speeds.
This is why OEMs increasingly treat optical assembly design as a core engineering consideration rather than an off-the-shelf component selection.
Our capabilities
We design and manufacture custom optical assemblies for OEM sorting equipment, supported by vertically integrated U.S. manufacturing.
Capabilities include multi-element doublet and triplet lens designs to correct for aberrations across demanding fields of view, image sensing and camera modification services, electro-optics component bonding and assembly, and Class 1000 and Class 100 cleanroom assembly cells for contamination-sensitive builds.
Every assembly is backed by sub-micron, traceable metrology and testing, ensuring assembly performance is measured and documented against specification requirements.
Contact our team of technical experts to discuss your application requirements: sales@torrentphotonics.com