MyAdvantech Registration

MyAdvantech is a personalized portal for Advantech customers. By becoming an Advantech member, you can receive latest product news, webinar invitations and special eStore offers.

Sign up today to get 24/7 quick access to your account information.

Advantech and Spingence Optimize Defect Detection with AI for Passive Component Manufacturing


Passive components are vital for electronic products, and their production is important in Taiwan’s electronics industry. For the past few years, Taiwan's passive component industry has seen active development for high-end applications such as electric vehicles, aerospace, and 5G equipment. A key contribution to this development has been the upgrade of Spingence’s AINavi manufacturing defect detection software in partnership with Advantech to improve production yields and reduce costs.

AOI Rule-Based Inspection Results in Three Major Blind Spots in Defect Detection

Inspection for external defects is a critical step in quality assurance before passive components are shipped. Jem Wu, Spingence’s Director of Business Development for Greater China, pointed out that passive components are mainly used for storing or releasing electricity, and external damages or other defects may cause these components to behave irregularly or fail, leading to dangerous situations such as short-circuits, fire, or explosion.

In the past, passive component manufacturers mostly conducted appearance defect detection using six-sided inspection machines based on automated optical inspection (AOI) technology, which compares optical images of items with prescribed rules to identify defects. "Such a rule-based inspection method is highly prone to missing defects, difficult system maintenance, and high overkill rates," said Wu.

To reduce the number of missed defects in inspections and get closer to the goal of zero missed defects, engineers generally use stringent parameters to define rules for identifying defects, which leads to a high overkill rate where flawless products may be misjudged as defective, resulting in increased cost and the need to double-check with human visual inspection. "Both the false alerts and the need for additional personnel are draining more production cost from our clients," said Wu.

Spingence Integrates AOI with AI to Maximize Defect Detection Performance

To solve issues with defect detection, Spingence launched the AINavi defect detection software to overcome some of the limitations of AOI technology with artificial intelligence. Since AI can learn the characteristics of defects from large amounts of historical data and identify defects through those characteristics, AI is superior at defect detection and can detect more defect types than AOI. This allows companies to reduce their risk of missing defects, while also reducing the overkill rate and the operating cost of hands-on maintenance.

"AOI and AI apply different principles to identify defects, and there is no way in which one is definitely better or worse,” Wu noted. “But by using both methods, they complement each other and maximize the performance of defect detection.” AOI inspection is based on rules, which are optimized for identifying defects related to measurements such as product length where a specific standard has been defined, and it does not require parameter adjustments. On the other hand, the strength of AI lies in identifying common features among many defects, making it more suitable for handling complex or ambiguous defects.

When supporting passive component manufacturers in implementing AINavi, Spingence provided standard APIs and also integrated existing inspection machines (6-sided) with TCP/HTTP, allowing production line processes to remain unchanged. Manufactured items would still pass through the six-sided inspection machines, except that the machines would send extra sets of images to the AINavi for AI system recognition. Then the results would be sent back to the six-sided inspection machines for follow-up processing. This means that workers do not need to learn new processes, which ensures continuous production line stability.

In addition to helping passive component manufacturers integrate their existing equipment, Spingence also takes into account the need for high efficiency within the manufacturing industry. AINavi runs on Advantech's fanless MIC-770 system with a graphics card, or it can run on the MIC-730 AI inference system. AINavi can thus inspect thousands of parts in just one minute, meeting the high-speed inspection needs of production lines by a wide margin.

According to Wu, two main factors contributed to their decision to collaborate with Advantech. The first is that Advantech hardware is highly diversified and lightweight, suiting the preferences and needs of a wide variety of clients. The second advantage is hardware stability. As industrial computers become loaded with graphics cards to run AI models, maintaining the integrity of the hardware system can be quite a challenge. As a reputable and trusted global hardware provider, Advantech provides some of the most stable and robust hardware on the market, making Advantech the best hardware partner for Spingence.

Three Major Benefits of AINavi Defect Detection Software: Reduction in Operating Time, Labor, and Cost

After successful cases introducing AINavi to passive component production lines, three major benefits have become clear. First is the saving of engineers' valuable time. In the past, engineers had to know every detail of a system to set parameters and rules to define particular types of defects, such as defining what constitutes a defective corner of a manufactured item. And if just one single defective corner was missed, they would have to adjust the AOI parameters. After the implementation of AINavi, however, engineers now only need to collect product images, draw circles around the defects, and leave the rest to the system as it learns the characteristics of the defects on its own. "AINavi’s training interface is simple and intuitive. Production line operators can learn how to operate it easily," said Wu.

The second benefit is the reduction in manpower required for visual inspection. sometimes being able to eliminate the whole manual double-check. For example, a manufacturer of passive automotive components originally had three operators for each six-sided inspection machine that were responsible for the loading, unloading, and double-checking. Since the introduction of AINavi, every three six-sided inspection machines only require one operator to handle both the loading and unloading. Such improvements not only reduce labor costs but also eliminate risks associated with labor shortages.

The third benefit is the reduction in unnecessary waste from the production process, which reduces production costs. A factory that produces passive components originally had an overkill rate of 4-5%. After the introduction of AINavi, many products that were originally classified as defective were recovered, while the overkill rate was reduced to 1%. This dropped production costs by 3-4% as a result.

Wu emphasized that since end customers’ expectations have been on the rise, the passive component industry can no longer solely rely on existing AOI equipment to avoid defects. Only by integrating AI technology to improve quality control capabilities and ensure that all shipped products meet customers’ needs will companies earn recognition and support from the market. This is true not just for the passive component industry, but for any manufacturing industry that is suffering from a high overkill rate with a need to allocate extra manpower for re-inspections. For manufacturers of ICs, fasteners, metal parts, etc., the introduction of AINavi will ensure product quality, offering opportunities to expand their markets by offering more refined products.