For the smart textile fabric inspection scene, China Science Visual Language and Huawei jointly create a smart textile fabric inspection joint solution. The scheme is based on the fabric defect detection algorithm developed by Shishyu and the quality inspection device polished for a long time. It has completed the adaptation of Huawei Centime series high-performance low-power edge computing unit and Sunsys MindSpore full-scene AI framework, giving full play to the integration ability of AI and production, comprehensively improving the quality inspection quality and efficiency of textile enterprises, and providing strong support for the intelligent and automatic upgrading of weaving enterprises.

With the rapid development of intelligent industry, traditional manufacturing industry's pain points are more obvious, with non-contact, high efficiency, low cost, high degree of automation advantages of detecting way become a "breakthrough" in the field of industrial flaw detection, traditional weaving industry using artificial quality testing, the efficiency of artificial visual inspection are easily influenced by emotional, physical, fatigue and environmental, The uncertainty of quality inspection efficiency will cause uncontrollable losses to enterprises. At the same time, enterprises have high employment costs and difficulties in recruitment management. Secondly, it is difficult to save and manage manual cloth inspection data, there is no standard format, and the upstream and downstream industrial chain is disconnected, and there is no evidence to check the defect detection, resulting in repeated resource consumption. Therefore, the traditional weaving industry urgently needs intelligent, digital and automatic detection methods.

Intelligent textile fabric inspection joint solution

Help textile enterprises reduce costs and increase efficiency in quality inspection

Huawei and zhongke joint solution based on visual language wisdom textile cloth inspection division depending on the language of light coupling neural network wide submillimeter level flaw detection technology, combined with huawei hardware performance advantage in end side and the ability of development and deployment support of deep learning algorithm, achieve far more than artificial fabric defects identification, data analysis, the machine monitoring, supervision over the performance and energy consumption, and other functions, A complete set of intelligent visual cloth inspection equipment integrating optical imaging device, edge computing device and textile big data management platform is constructed to serve the automatic production line upgrading of the whole process of fabric quality inspection.

Stabilize quality inspection, reduce costs and increase efficiency. The intelligent cloth inspection solution brings the possibility to the unattended quality inspection process. The AI visual cloth inspection machine can be free from the influence of emotion, body, fatigue and environment for 24 hours all day, and carry out the cloth defect detection at a stable speed 3-5 times that of manual labor, which improves quality inspection and reduces cost at the same time.

Data traceability and standardization of defects. The quality test results and corresponding rating information of each cloth roll are recorded in the whole process, which can be stored digitally and completely, with perfect traceability, effectively avoiding unnecessary disputes and realizing the standardization of textile defects.

Data analysis, help enterprise management level upgrade. Make intelligent big data analysis on the whole process data stored permanently, master the key information elements of enterprise production, timely adjust the state of production factors to improve production efficiency, and achieve high-level fine management.

All domestic intelligent equipment, the whole process is autonomous and controllable. Self-developed domestic solutions, reduce the potential risks brought by the macro environment, the whole process of autonomy and control, high flexibility, support deep customization, provide exclusive production line empowerment solutions for enterprises.

Keep learning, keep evolving. The system has excellent ability of evolution and expansion, continuous self-learning in the application process, and can be extended to support new fabric and new types of defects.

See language intelligent cloth inspection device to complete Huawei mainstream adaptation

Through the Atlas computing hardware and MindsporeAI framework dual certification

It is worth mentioning that at present, the intelligent fabric inspection equipment of Visual Language has completed the mainstream adaptation of Huawei and passed the dual certification of Atlas computing hardware and Mindspore AI framework. The certification results fully verify the advantages of high integrity, openness, compatibility and maturity of visual language products. At the same time, it further strengthens the characteristics of independent innovation, safety, reliability, stability and reliability of visual language products.

Up to now, the scheme has been implemented in many textile head enterprises, helping customers greatly improve the detection rate, accuracy and detection speed of fabric defects, improving the yield of textile products, alleviating problems such as difficulty in ensuring the accuracy of manual eye inspection, increasing the efficiency by about 50% and saving the cost by 70%. The application effect is well received by customers.

Hai'an County Qinfeng Chemical Fiber Co., Ltd. specializes in production and sales:polypropylene staple fiber, Polyester Staple Fiber, functional polypropylene short fiber, functional polyester staple fiber, hydrophilic polyester staple fiber,  polypropylene, PP staple fiber, polyester staple fiber.