Technology & research

From imaging to understanding.

Connect optical and internal imaging with vision algorithms and industrial LLMs, bringing together acquisition, geometric measurement, intelligent recognition and knowledge processing for manufacturing.

01Visible light

02X-ray

03Vision AI

04Industrial LLMs

VISIBLE-LIGHT INSPECTION

Visible-light inspection

Reveal surface detail and geometric form.

Capture surface detail with microscopy and high-resolution imaging, and acquire height and shape with profile sensing and line chromatic confocal measurement. Coordinate illumination, optics, cameras, sensors and algorithms to connect 2D appearance with 3D geometry.

Microscopy: Resolve fine structures with microscope optics, matching magnification, working distance and illumination. Focal-plane acquisition and image analysis extract edges, positions and geometry.
Ultra-high-resolution cameras: Retain textures, edges and small-defect information through high-pixel-count acquisition and fine sampling. Coordinate field of view, magnification and synchronized exposure for coverage and detail.
3D profile sensors: Acquire cross-sectional profiles by laser triangulation, combining translation or rotation to build point clouds and height maps. Analyze steps, grooves, flatness and profile changes.
Line chromatic confocal: Relate wavelength to focal distance through chromatic optics and recover surface height from the return spectrum. Acquire heights across a measurement line and reconstruct steps and 3D surface shape by scanning.

X-RAY INSPECTION

X-ray internal inspection

Look beneath the surface of the package.

Build internal imaging around microfocus X-ray sources, high-resolution detectors and precision motion, capturing transmission differences in materials and structures. Combine digital processing, geometric measurement and automated acquisition for radiography and structural analysis.

Microfocus X-ray sources: Combine a small focal spot with geometric magnification to observe fine internal structures. Coordinate source, sample and detector positions for the required field and viewing angle.
High-resolution detectors: Convert transmitted X-ray signals into digital images that preserve edges and intensity differences. Combine exposure, sampling and noise processing to support image analysis.
Image & structure analysis: Filter projections, extract regions, assess positions and measure geometry. Represent internal wires, connections and package outlines as image features for analysis.
Motion & automated acquisition: Coordinate positioning, handling and capture timing for multiple views and inspection sequences. Link feeding, image records and sorting with equipment actions and inspection data.

INDUSTRIAL VISION

Industrial vision & intelligent sensing

Understand people, parts and equipment on site.

Connect image data, model training and inference to turn visual features into objects, locations, states and anomalies. Combine recognition models with process rules to link perception and task decisions in manufacturing.

Object detection: Identify people, parts and equipment, returning classes and locations. Organize target regions as inputs for counting, region decisions and process analysis.
State recognition: Use features, spatial relationships and process rules to describe presence, absence, assembly errors and action states. Combine recognition with business conditions for task decisions.
Anomaly detection: Analyze defects and normal states to identify appearance differences and changes on site. Combine thresholds, rules and model outputs to organize anomaly classes and review information.
Training & inference: Organize import, annotation, training and evaluation, with model export and algorithm workflows. Connect inference results with business systems, linking acquisition, algorithms and presentation.

INDUSTRIAL LLM / SMARTINSIGHT

Industrial LLMs & intelligent applications

Put manufacturing knowledge to work.

Build language-model applications around manufacturing data and enterprise knowledge. SmartInsight focuses on structured business data, combining shared datasets, natural-language queries, interactive reports and analysis for production, quality and business management.

Enterprise knowledge: Organize process documents, work instructions, equipment references and historical issues into searchable knowledge. Preserve sources and relationships to provide enterprise context.
Retrieval-augmented generation: Retrieve relevant materials and organize document excerpts as model context to generate source-grounded answers. Link enterprise retrieval with language generation, connecting questions, evidence and responses.
Manufacturing data analysis: Link inspection records, batches and historical issues to organize anomalies, leads and summaries. Combine structured manufacturing data with text knowledge for review and traceability.
Models & business integration: Organize model calls, data interfaces and permissions for each task, connecting retrieval, responses and workflows. Retain references, original records and human review for business use.