A bone splinter in a chicken fillet. A plastic fragment in a ready meal. A piece of cartilage that no X-ray system in the world can distinguish from the meat around it — because its density is nearly identical to the product. These so-called density-neutral foreign bodies are the largest unsolved challenge in food safety. They trigger recalls that cost companies an average of USD 10 million in direct costs, and up to 55 percent of consumers switch brands afterward.
There's a misconception still common in many QA departments: "X-ray finds everything." It doesn't. Metal detectors don't find glass, plastic, or bone. Conventional X-ray systems only detect foreign bodies whose density differs clearly from the product — hair, insects, feathers, and soft bone all slip through. And manual visual inspection is simply impossible at modern production speeds, with multihead weighers running over 200 cycles per minute.
The answer lies in combining intelligent technologies with a holistic inspection strategy for foreign body detection in food products — from raw material to the ready-to-ship product.
The decisive physical variable is density difference: the greater the density gap between a foreign body and the product, the easier it is to catch by X-ray. A metal fragment in bread dough? No problem. A glass shard in a can? Reliably detectable. But there are contaminants that conventional technology systematically fails to catch, which is exactly where dedicated foreign body contamination detection becomes essential:
Anyone who relies solely on conventional technology accepts a residual risk with these materials.
X-ray systems work on the principle of density differentiation: areas of higher density absorb more radiation and create contrast in the image. That works well for metal, glass, or stone — but barely at all when the density of the foreign body and the product are similar. Add to that interference factors such as product overlap, packaging seams, and clip closures, which can mask foreign-body signals. Conventional rule-based software either responds with more false rejects or has to lower sensitivity, risking that real foreign bodies slip through.
Metal detectors are proven for their purpose but detect metals exclusively — glass, stone, plastic, bone, and wood pass through undetected. And manual visual inspection reaches its limits at production speeds above 200 cycles per minute: human attention drops noticeably after just 20 to 30 minutes of monotonous visual checking.
One way to narrow the detection gap for density-neutral foreign bodies is AI-assisted image analysis — used at RaymanTech under the name isiray, one of several technologies in the product portfolio alongside metal detection, X-ray inspection, and optical sorting. Rather than working with rigid thresholds, the approach draws on numerous specialized algorithms, including shape recognition, which identifies foreign bodies by their geometric signature rather than density contrast alone. Trained on real inspection images, the AI can pick up on finer density and texture variations that a rule-based evaluation would miss.
Another building block is dynamic masking: packaging edges, clip closures, and weld seams are detected in real time and excluded from the analysis instead of relying on fixed zones. This can reduce false alarms at packaging seams and improve detection near the packaging edge. Importantly for setting expectations: even AI-assisted image analysis does not reliably catch every density-neutral foreign body. Soft plastics in particular remain a known limitation — for very soft, thin-walled plastics or organic materials, optical systems or combined solutions are often the more reliable complement.
No single technology covers the entire foreign-body spectrum. That's why RaymanTech takes a consultative approach with its Whole Chain Inspection Solution (WCIS): the entire production line is analyzed to define the right combination and placement of different inspection technologies for foreign body detection in food — matched to the product, risk, and regulatory requirements.
A typical setup combines optical sorting at goods-in, metal detectors at critical processing points, X-ray inspection (optionally with AI-assisted image analysis) at final inspection, plus complementary visual inspection. What one system misses, the next is ideally positioned to catch. RaymanTech also offers combo systems that combine X-ray and optical inspection in a single machine, covering a broader foreign-body spectrum in one pass.
In practice, these technologies complement each other: X-ray covers embedded foreign bodies that aren't visible optically; optical and hyperspectral systems cover low-density and organic foreign bodies that X-ray fails to catch.
Standards such as HACCP, IFS Food, and BRCGS require a documented foreign object detection procedure at critical control points, plus regular validation of the systems in use. These standards form the backbone of foreign object detection food safety programs across the industry. Increasingly, major retail chains also enforce a "zero glass policy" that goes beyond metal detection alone and requires proof of effective X-ray inspection. Companies that fail to meet these requirements risk losing retail listings.
Investment costs for foreign body detection equipment rise significantly from metal detectors through X-ray and optical systems to hyperspectral solutions. Beyond the purchase price, what matters most for the cost-effectiveness assessment is total cost of ownership: avoided recall costs, reduced false rejects, audit compliance, and retention of retail listings. RaymanTech positions itself with a TCO that, according to its own figures, runs 30 to 50 percent below comparable European providers, plus local service from a test center in Germany where systems can be trialed on customers' own products.
The terms are used interchangeably and describe the technical process of identifying and rejecting unwanted particles or objects in products.
X-ray inspection reliably detects foreign bodies with a clear density difference from the product: metals, glass, stone, ceramics, calcified bone, and hard plastics. Density-neutral foreign bodies such as cartilage, soft bone, or thin plastics, as well as organic contaminants, frequently slip past conventional systems.
Partially. Hard plastics and calcified bone can generally be detected well given sufficient density contrast. Soft plastics, wood, and non-calcified bone, however, are not reliably detected with conventional X-ray technology. AI-assisted image analysis can improve detection in some of these cases but is no complete guarantee — for wood, paper, and many plastic variants, optical or hyperspectral systems usually remain the more reliable choice.
A combination of approaches makes sense: AI-assisted image analysis, which can additionally catch some density-neutral contaminants; optical inspection or combo systems for foreign bodies on the product surface; and a WCIS strategy that combines multiple technologies at different points along the production line.
No. Industrial X-ray inspection systems operate at extremely low, legally certified radiation doses. With proper operation, there is no radiation risk to operating staff.
Costs vary significantly by technology: metal detectors sit at the lower end, X-ray systems in the mid-to-upper range, and optical and hyperspectral systems in the mid-to-upper range as well. What determines ROI is not the purchase price alone but the total cost of ownership, including maintenance, false-reject costs, and audit compliance.
