碳纖維布與複合片材 AI-AOI 視覺檢測系統

AI-AOI Visual Inspection System for Carbon Fiber Fabric and Composite Sheets

岳揚智控（AI Machine Vision）憑藉榮獲 愛迪生獎（Edison Awards）金牌 的 [ 紡織布料檢驗 AI 視覺檢測技術 ]，正式跨足複合材料領域。針對碳纖維織布（Carbon Fiber Fabric）極高的反射性、複雜紋理及零容忍的品質要求，我們提供一站式 AI+AOI 自動化檢測方案。從碳纖原布、預浸材到複合成品，即時攔截微米級瑕疵，助您從傳統人工目檢或傳統 AOI 瑕疵偵測，轉型為 AI-AOI 視覺數據驅動的智能製造，迎接 AI Agent (AI 代理人) 時代來臨，紮下世界級智慧企業的基礎。

Building on its Edison Gold Award-Winning AI vision technology for textile inspection, AI Machine Vision Corp. (AIMV) is officially expanding into the advanced composite materials sector. To address the extreme reflectivity, complex woven textures, and zero-tolerance quality requirements of carbon fiber fabric, we deliver a total AI+AOI automated inspection solution.

Spanning dry carbon fiber fabric, prepreg, and finished composite laminates, our system detects micron-scale defects in real time. We enable your transition from manual visual inspection or traditional AOI parameter tuning to AI-AOI visual data-driven smart manufacturing—laying a world-class foundation for modern intelligent enterprises in the era of AI Agents.

1. 全光譜影像感測：全頻域與多角度光源配置，有效克服碳纖維的高反射與（Black-on-Black）影像低對比問題。
2. 強化的深度學習算法：超越傳統 AOI，能精準識別非固定形狀的瑕疵，檢出率高達 95% 以上。
3. 滾動式學習：具備「人機協作」滾動學習機制，AI 模型訓練期間不影響產能下，可使系統越檢越聰明，降低（誤殺／誤放）率。
4. 高速線上檢測：檢測速度最高 （60碼／分）的檢驗速度，無縫整合進 Roll-to-Roll ／Roll-to-Sheet／Roll-to-Sheet 生產製程。
5. 解決職業傷害：碳纖維檢驗需長時間高強度照明，人工目檢易導致視網膜傷害，落實 ESG 企業社會責任。
6. 數位化履歷驗報：每一捲布料／片材，均產出專屬的數位品質地圖（Defect Map），記錄瑕疵類別、位置與大小。這些數據可回溯至生產製程，協助研發團隊優化編織與浸漬工藝，解決生產工藝的品檢「數位斷層」資訊。
7. 系統模組化配置：不論是平織、斜織、單向布，系統ＡＩ演算法皆採模組化設計，可根據現場幅寬（最高支援 2,000mm）與製程需求彈性調整。

Key Features & Technological Advantages:

1. Full-Spectrum Image Sensing: Configured with full-spectrum and multi-angle lighting to effectively overcome high reflectivity and low contrast ("black-on-black") imaging challenges inherent to carbon fiber.
2. Enhanced Deep Learning Algorithms: Transcending traditional AOI, the system accurately identifies irregular and non-fixed defect shapes, achieving a defect detection rate exceeding 95%.
3. Continuous Learning Mechanism: Features a human-in-the-loop rolling learning mechanism that enables continuous model evolution without interrupting production throughput, thereby minimizing both false call and overkill/escape rates.
4. High-Speed In-Line Inspection: Delivers inspection speeds up to 60 yards per minute, seamlessly integrating into Roll-to-Roll (R2R) and Roll-to-Sheet (R2S) manufacturing processes.
5. Mitigating Occupational Hazards: Manual inspection requires prolonged exposure to high-intensity lighting, which poses severe risks of retinal injury. Automating this process fulfills key ESG corporate social responsibility standards.
6. Digital Inspection Traceability: Generates a dedicated digital quality map (Defect Map) for every roll or sheet, recording defect classification, coordinates, and size. This data traces back to production processes, empowering R&D teams to optimize weaving and impregnation recipes while closing the "digital gap" in quality control.
7. Modular System Architecture: Designed with modular AI algorithms that adapt to plain, twill, or unidirectional (UD) weaves. The architecture flexibly scales based on on-site web widths (supporting up to 2,000 mm) and specific line requirements.

碳纖維預浸材（Prepreg）採用高階 AI 視覺檢測

用全頻域與多角度結構光源，挑戰黑對黑的視覺極限，實現樹脂含量與結構監控在預浸材的含浸品質，樹脂的均勻度與纖維的平整度決定了最終複合材料的力學強度。

1. 攻克「黑對黑 (Black-on-Black)」檢測難題

預浸材表面充滿黑色樹脂與黑色碳纖維，傳統視覺極難分辨。

• 多角度光源技術：透過特定波長的 LED 陣列光源，突顯樹脂流動痕跡與纖維絲束的微觀起伏，使隱藏在深色樹脂下的斷經、亂緯無所遁形。
• 動態補償（HDR）：確保在高度反射的樹脂表面與深色吸收光的纖維區域，皆能獲得高對比度的清晰影像。

2. 預浸材專屬瑕疵檢測項目

針對預浸料生產線，精確識別以下核心瑕疵：

• 塗佈膠厚：同步採用雷射測厚儀量測與ＡＩ視覺時時監控。
• 捲曲折痕：在含浸滾壓過程中，即時攔截因張力不均產生的微小皺褶。
• 膠塊雜質：識別未融化的樹脂塊或離型紙/膜脫落碎片。
• 經緯密度：嚴密監控經緯向間距，確保結構強度符合航太與車用規範。

3. 實時寬度與邊緣監控

• 高精度尺寸量測： 系統即時監控預浸材的有效幅寬及邊緣整齊度，確保收捲過程不偏移，降低後續裁切損耗。
• 離型紙/背襯膜檢查：自動確認背襯材料是否平整覆蓋，防止黏合失敗導致的整捲報廢。

4. 製程優化數據鏈結岳揚智控的系統不只是「篩選瑕疵」，更是「製程診斷器」

• 趨勢分析：當系統偵測到特定位置反覆出現相同瑕疵，將自動觸發報警，提示操作員檢查。
• 數位地圖：自動產出瑕疵座標地圖資訊，提供自動裁切機自動避開瑕疵區塊，最大化材料利用率。

5. 技術參數摘要：

• 最小瑕疵：可達 0.45mm（依檢測需求，可更微細到0.1mm）。
• 系統整合：支援 PLC 通訊與工廠 MES 系統對接，落實生產自動化。

Advanced AI Vision Inspection for Carbon Fiber Prepreg

Tackling the "Black-on-Black" Visual Limits with Full-Spectrum and Multi-Angle Structured Lighting for Resin Content and Structural Monitoring

In prepreg impregnation quality, resin uniformity and fiber flatness dictate the mechanical strength of the final composite material.

1. Overcoming the "Black-on-Black" Inspection Challenge

The prepreg surface is dominated by black resin and black carbon fibers, making differentiation extremely difficult for conventional vision systems.

• Multi-Angle Lighting Technology: Utilizes specific wavelength LED array lighting to highlight resin flow marks and microscopic fiber bundle undulations—exposing broken warps and misaligned wefts hidden beneath dark resin.

• Dynamic High Dynamic Range (HDR) Compensation: Ensures clear, high-contrast imagery on both highly reflective resin surfaces and light-absorbing dark fiber zones.

2. Specialized Defect Inspection Scope for Prepreg

Engineered specifically for prepreg production lines to precisely identify core defects:

• Coating Thickness: Synchronizes laser thickness gauge measurements with continuous real-time AI vision monitoring.

• Winding Creases & Wrinkles: Intercepts micro-wrinkles caused by tension imbalance in real time during the impregnation rolling process.

• Resin Clumping & Impurities: Detects unmelted resin lumps or flaked release paper/film debris.

• Warp and Weft Density: Rigorously monitors warp and weft spacing to guarantee structural strength compliance with aerospace and automotive standards.

3. Real-Time Width and Edge Monitoring

• High-Precision Dimensional Measurement: Continuously monitors effective prepreg width and edge alignment to prevent misalignment during winding and reduce subsequent cutting waste.

• Release Paper / Backing Film Inspection: Automatically verifies flat backing material coverage, preventing entire roll scrap caused by bonding failures.

4. Process Optimization Data Linkage: A "Process Diagnostic Tool" Beyond Quality Screening

• Trend Analysis: Automatically triggers alerts when recurring defects appear at specific locations, prompting operators to inspect machinery.

• Digital Mapping: Generates automated defect coordinate maps to guide automated cutting machines in bypassing defect zones, maximizing material utilization.

5. Key Technical Specifications Summary

• Minimum Defect Detection Size: Down to 0.45 mm (configurable down to 0.1 mm based on inspection needs).

• System Integration: Supports PLC communication and factory MES system interfacing to enable full factory automation.

複合材料的品質，值得世界級的 AI 守護

用愛迪生金獎級的ＡＩ視覺辨識技術，守護你的碳纖維布與複合片材品質。

[立即預約諮詢]

E-Mail : service@ai-machinevision.com

 TEL : 04 - 2491 2151