OEM technology partnerships
DIRATECH works with machine builders, system integrators and X-ray equipment suppliers to develop inspection capabilities for their customers. Our contribution can include inspection software, detection modules, hardware selection and integration engineering, with the scope defined around your application.
In products with mixed densities, uneven textures or overlapping pieces, contaminants can be difficult to distinguish from the surrounding product on an X-ray image. Even metal, glass or stone can be challenging to detect in these conditions.
DIRATECH combines X-ray imaging with AI-based analysis to help distinguish contaminants from the product’s natural structure. We work with food producers and equipment manufacturers to evaluate these challenging applications and develop an inspection approach suited to the product.
Testing with representative samples establishes what can be detected and under which conditions, including the balance between detection performance and false rejects.
DIRATECH conducted a comparative detection study using 500 g double-layer pistachio bags and a reference set of customer-collected glass and stone contaminants. The test set focused on the most difficult-to-detect pieces, selected to represent a demanding real-world inspection challenge.
The benchmark system was a leading globally established X-ray inspection brand. To ensure a fair comparison, both systems were set up and calibrated by their respective engineers under routine operating conditions. Only contaminant samples demonstrating repeatable detectability in preliminary 20-pass verification testing on both systems were selected for the comparative dataset. Final detection rates were then calculated across the full comparative test sequence.
| Contaminant type | DIRATECH detection rate | Leading global X-ray system |
|---|---|---|
| Glass fragments | 96% | 49% |
| Stones | 100% | 60% |
Results are specific to the defined product, reference contaminants, test sequence and operating conditions; performance may vary by application.
The results show a substantial detection advantage for DIRATECH on challenging contaminants embedded in complex, double-layered pistachio products.
DIRATECH’s performance is enabled by its patent-pending Hybrid Detection Engine, which combines deep learning, material separation, and advanced image processing. Rather than relying on a single inspection approach, the system evaluates product images through complementary detection layers designed to distinguish genuine foreign-material signals from normal product variation.
Using standard DIRATECH settings and a rapid commissioning procedure, the system achieved these true-positive results while maintaining a 0.3% false-reject rate. This rate was measured across 10,000 highly heterogeneous images of double-layered pistachios collected on a real production line.
The platform supports site-specific continuous learning and image-data retraining, enabling further optimization of false-reject performance as validated production data are accumulated.
For machine builders and integrators introducing X-ray inspection, we can contribute detection technology and engineering support to help define the inspection setup, select components and develop the integration into your machine.
For established X-ray equipment suppliers, we start with a specific application or customer requirement. Together, we assess where DIRATECH technology could complement your current platform and what would be needed to evaluate and integrate it.
The first question is what your customer needs to inspect. We use that requirement to define the technical scope, assess compatibility and agree what should be demonstrated before committing to an integration project.
Image analysis and detection logic adapted to the product and inspection criteria.
Support in defining the X-ray source, detector and acquisition setup, with inspection components configured for the application.
Support for machine interfaces and joint development, with compatibility and responsibilities agreed during the technical review.
Sample evaluation, application-specific model development and agreed support for the implemented technology.
Share the product and packaging, target contaminants or defects, required throughput and relevant details of your platform.
We define representative samples, test conditions and acceptance criteria. The evaluation setup is agreed according to the application and available equipment.
We assess detection, false rejects and speed where relevant, then define interfaces, engineering work and responsibilities.
Development, software licensing and support are scoped for the project. The route to deployment follows the agreed technical and commercial milestones.
Xensio is our development programme for challenging food inspection applications using multi-view imaging, 3D reconstruction and AI. We are inviting food producers and technology partners to discuss pilot projects.
Tell us about your platform, the product you need to inspect and the result your customer needs. We will review your enquiry and contact you to discuss the next technical step.
Enquiry received
Thank you. Your enquiry has been received. Our team will review your application and contact you to discuss the next technical step.
