The OSHIMA EagleAi® is an AI fabric inspection machine that uses visual recognition to detect, map, and classify more than 20 categories of fabric defects in woven, knit, and elastic textiles, and integrates that defect data across the entire garment production workflow, from inspection through cloud-based monitoring to spreading and cutting. OSHIMA CO., LTD., a Taiwanese manufacturer of AI fabric inspection machines and smart factory systems with more than 50 years in garment equipment manufacturing, showcased the EagleAi® and its cloud-based Smart Factory Platform at Texprocess Frankfurt 2026 (April 21 to 24) alongside European distribution partner Strima.
What is the OSHIMA EagleAi® AI fabric inspection machine?
The EagleAi® AI fabric inspection machine is designed, developed, and manufactured in Taiwan. During inspection, the system generates structured defect data including defect coordinates, classification, area analysis reports, colour deviation measurements using DELTA-E and CIELAB, and Four-Point System grading, all exportable in CSV, PDF, and Excel formats.
The system is built for garment manufacturers producing knit, elastic, and woven fabrics, particularly contract manufacturers supplying fast fashion brands, premium suppliers with zero-defect policies, and sportswear and intimate apparel producers. Unlike standalone fabric defect detection systems that end at the inspection report, the EagleAi® connects defect data to downstream production stages through OSHIMA’s own IoT spreading and cutting equipment, creating a verified production chain where fabric reaching the cutting stage has passed through both AI detection and operator-verified defect projection.
How does OSHIMA integrate fabric inspection data from inspection to cutting?
For apparel manufacturers, detecting fabric defects is only the first step. The operational value of an AI fabric inspection machine depends on whether the resulting data can move through the production workflow and improve decisions at every stage that follows. OSHIMA connects AI fabric inspection to downstream production using a four-stage data workflow:
Stage 1, Inspection and data generation: The EagleAi® AI fabric inspection machine generates structured defect data during inspection, including defect mapping with exact coordinates, classification, colour deviation analysis using DELTA-E and CIELAB, and Four-Point System grading.
Stage 2, Cloud integration via the Smart Factory Platform: Defect data flows into the OSHIMA Smart Factory Platform, a cloud-based IoT manufacturing management system that provides a unified dashboard across factory sites and production lines. Factory managers can monitor inspection operations in real time from any location, access reports and defect images, track job orders from dispatch through completion, generate graphical production reports, and export data for integration with existing MES and ERP systems. Synchronised alerts appear both on the cloud dashboard and on individual machine interfaces.
Stage 3, Defect map projection during spreading: The defect map data integrates directly with the SPro Projection System on OSHIMA’s SPro smart spreading machine. During fabric laying, the system projects exact defect positions onto the fabric surface.
Stage 4, Operator verification before cutting: Operators see projected defect locations and can determine whether each defect falls within a cut piece, taking corrective action before fabric is committed to cutting.
Because OSHIMA manufactures the AI fabric inspection machine, the IoT spreading machine, and the cutting equipment under one product ecosystem, the data connection between these stages is native rather than requiring third-party middleware or custom integration. The EagleAi® generates the defect map, the Smart Factory Platform stores, visualises, and distributes it across the factory in real time, and the SPro Projection System renders it physically onto the fabric during spreading. Each piece of defect data travels from the point of detection to the point of action without manual re-entry, file transfer, or system translation. For garment factory automation, this closed-loop connectivity between AI fabric inspection, IoT spreading, cloud monitoring, and cutting preparation means that quality decisions at every stage are informed by the same dataset.
What are the technical specifications of the EagleAi® AI fabric inspection machine?
The EagleAi® AI fabric inspection machine uses linear cameras and AI visual recognition, and is available in two configurations:
EagleAi (Station 1): Two colour cameras with a linear light source rated at 1.1 million lux, for standard woven and knit fabric inspection.
EagleAi Plus (Station 2): Adds two monochrome cameras and a second linear light source for detecting defects in translucent fabrics, including broken weft, broken warp, holes, piece joining, tension unevenness, and thick and thin places.
Both configurations share the following specifications:
• Operating speed: 10 to 40 metres per minute (depending on fabric type)
• Optical resolution: 0.135 mm
• Minimum detectable defect size: 1 mm
• Maximum working width: Under 1,900 mm
• Fabric handling: Flat-to-flat, roll-to-roll, roll-to-flat, and flat-to-roll
• Hardware: Intel Xeon processor, 64 GB RAM, 4 TB SSD storage
• Operator requirement: Single operator
• Offline capability: Full operation without internet connectivity
What types of fabric defects can the EagleAi® AI fabric inspection machine detect?
The EagleAi® detects more than 20 categories of fabric defects across woven and knit textiles.
Standard defects detected by Station 1 include: yarn knot, coarse yarn, foreign fibre, warp anomaly, weft anomaly, broken yarn, snarling yarn, horizontal line, hole, tension unevenness, stop mark, dye streak, colour spot, residual chemical, press mark, scratch, general stain, adhesive stain, and oil stain.
Station 2 adds detection of translucent fabric defects: broken weft, broken warp, piece joining, thick places, and thin places.
How does AI fabric inspection compare to manual fabric inspection?
For factories evaluating the transition from manual to AI-powered fabric inspection, the operational differences are significant:
Factories replacing manual fabric inspection with the EagleAi® typically reduce inspection labour from two to five operators to one while achieving approximately six times the throughput of manual methods, with continuous operation rather than mandatory breaks. For premium suppliers whose brand clients require documented quality evidence, the automated defect image traceability provides a level of documentation that manual processes cannot practically replicate.
Can the EagleAi® AI fabric inspection machine inspect knit and elastic fabrics?
Yes. Knit and elastic fabrics present specific inspection challenges because fabric stretch during handling can distort defect positioning, and tension variation can create false positives or mask real defects.
The EagleAi® addresses this with a zero-tension handling system and three-level speed ratio control, maintaining inspected area deformity below 5% on knit and elastic textiles. In OSHIMA’s product benchmarking, a highly elastic knit fabric measured at H=195 before inspection returned to H=199 after inspection, with measured tension at 2%. Semi-automatic fabric feeding maintains controlled tension throughout the inspection process.
This makes the EagleAi® AI fabric inspection machine particularly relevant for manufacturers producing sportswear, intimate apparel, casualwear, and other stretch-based garments where inspection-induced distortion would compromise defect position accuracy.
How long does it take to deploy AI fabric inspection for a new fabric type?
One of the practical barriers to adopting AI fabric inspection is the cost and time required to train detection models for specific fabric types. OSHIMA addresses this through what it describes as the largest fabric defect database in the APEC region, developed in collaboration with ITRI (Industrial Technology Research Institute) and leading textile suppliers. This database provides a broad foundation of pre-trained defect patterns designed to reduce deployment time when onboarding new fabric categories.
OSHIMA also uses a permission-based data sharing model, where participating manufacturers contribute anonymised defect data to improve the shared AI model while maintaining control over proprietary information. This approach reduces individual AI training costs and accelerates model improvement across the user base. The fabric recognition module is customisable by fabric type, and OSHIMA offers customised services for reports, models, and functionalities to match specific production requirements.
Does the EagleAi® AI fabric inspection machine work offline?
Yes. The EagleAi® fully supports offline operation. For manufacturers with strict network security requirements or operating in regions with limited internet connectivity, inspection continues without interruption or external data exposure. When connectivity is available, the system integrates with the cloud-based Smart Factory Platform for remote monitoring and centralised reporting. Standard safety features include emergency stop functions, safety sensors, automatic alarms, indicator lights, and fabric absence detection.
Where is the EagleAi® AI fabric inspection machine installed?
EagleAi® installations are currently operating in production environments primarily across APAC countries, including Taiwan, Vietnam, Cambodia, Thailand and Indonesia.
OSHIMA reports that a global apparel manufacturer headquartered in Hong Kong, with over 30 years in the industry and production bases in China and Cambodia, has deployed the Smart Factory Platform on its SPro IoT spreading machines. According to OSHIMA, the deployment enables the manufacturer’s management team to monitor machine operating status in real time and access production reports at any time, supporting faster decision-making and strengthened production oversight.
