This article is reposted by Xinlaifu New Materials Co., Ltd. The original author retains all copyright; the original link is provided at the end under "Sources".
The Newlife - Radbar team presents the following translated industry intelligence.
Interview with Wang Wei of Hik Ruiying: Century-Old X-Ray Technology is Being Redefined by AI
Enabling machines to "see and see clearly," allowing equipment to understand, judge, and continuously evolve on its own.
If we turn the clock back ten years, very few people would have viewed industrial X-ray inspection (X-Ray) as a highly competitive track with rapid technological development.
The discovery of X-rays dates back over a hundred years. In the industrial field, it has long been applied in scenarios such as food foreign object detection, PCBA solder joint inspection, and industrial flaw detection. Specifically regarding equipment, core components like the X-ray source, detectors, and mechanical structures have long possessed mature technological reserves. Overseas leading manufacturers have even been deeply engaged in this field for decades, making it a standard mature track.
However, Wang Wei, President of Hik Ruiying, offers a different judgment: "Industrial X-Ray has just begun, and the entire industry is still in its early stages. If you look back, it's a red ocean; if you look forward, it's a blue ocean."
At this year's CIIF (China International Industry Fair), Hik Ruiying officially released its high-precision 3D AXI inspection equipment for semiconductor packaging, advancing inspection capabilities to the sub-micron scale. While maintaining reliability and consistency, the equipment can achieve sub-micron measurement, 3D image reconstruction, and defect analysis, thereby entering higher-threshold markets such as advanced packaging and laboratory failure analysis.
From millimeter-level security inspection equipment to micrometer-level industrial inspection, and now to sub-micron-level measurement; accompanied by higher precision, stronger penetration, and higher defect detection capabilities. In just a few years, Hik Ruiying has carved out its own place and established a firm foothold in the X-ray inspection field with efficient R&D speed and continuous project implementation.
But what truly attracts attention is the new "variable" in the X-Ray track that Hik Ruiying is going all-in on—AI.
For example, taking the unverified Apple foldable phone iPhone Duo as an instance, the hinge technology of foldable phones has once again been pushed into the spotlight. A foldable phone's internal hinge may contain hundreds of零部件 (components). In the past, manual inspection with a magnifying glass could check the surface, but it was difficult to stably judge whether there were internal defects or foreign objects, directly affecting product yield.
A few years ago, Hik Ruiying began using X-Ray combined with AI technology for automatic hinge inspection. Wang Wei told 36Kr that at the time, the hinge was a new process facing pain points such as the outflow of various defects like foreign objects, cracks, voids, and missing assemblies. Through 100% AI X-ray inspection, the defect rate dropped by an order of magnitude.
In addition, in meat processing, X-Ray was originally mainly responsible for detecting residual bones and foreign objects. After Hik Ruiying integrated image AI algorithms and large model capabilities into the equipment, it further achieved fat-rate detection. In scenarios such as complete laptop inspections, internal springs of eyebrow pencils, and automotive零部件 (components), X-Ray, which rarely entered production lines in the past, is also beginning to gain new application opportunities due to the emergence of AI.
AI is becoming the new "winning hand" in the industrial X-Ray track.
Founded in 2020, Hik Ruiying officially entered industrial inspection in 2021. Today, its industrial inspection business covers multiple fields such as electronic manufacturing and food. Backed by Hikvision's powerful technological R&D system, it has quickly risen to a first-tier player in the industrial X-ray field, entering the top three domestically in multiple segmented tracks.
People's perception of Hikvision often remains stuck on "the global number one in security," but the reality is far from this. Since 2016, Hikvision has quietly initiated a decade-long strategic transformation. Over the past ten years, Hikvision has continuously spilled its capabilities in IoT perception and AI into industrial fields such as semiconductors, automobiles, and 3C. Its "Banyan Business Model" has built eight major innovative business segments including Hikrobot, Hikmicro, and Hikvision Automotive Electronics, completing the strategic upgrade from a security equipment and solutions provider to an AIoT technology company.
Hik Ruiying is precisely one of the key business segments of the "Banyan Model."
Wang Wei describes future X-Ray equipment as "intelligent equipment robots." In the future, equipment robots will be able to automatically discover defects and understand them; they will also automatically adjust parameters and further use inspection data to inversely influence front-end processes.
"The future will definitely be intelligent. Today's highest requirements may be tomorrow's lowest standards," Wang Wei said.
From enabling machines to "see and see clearly" to attempting to let equipment understand, judge, and even continuously evolve on its own, the path Hik Ruiying has taken in X-Ray also provides us with a close-up window to observe this emerging industrial leader under the Hikvision umbrella.
The following is the Q&A transcript between 36Kr and Wang Wei, President of Hik Ruiying.
I. "There Are Few True Global Players; You Can Count Them on Two Hands"
36Kr: What is the core technical breakthrough of the sub-micron high-precision 3D AXI inspection equipment for semiconductor packaging released this time?
Wang Wei: We call this product 3D AXI internally. Its core achievement is realizing sub-micron X-Ray measurement under conditions of high reliability and high consistency. Sub-micron means less than one micrometer, entering the nanometer scale, while also completing 3D image reconstruction and defect analysis.
It has a relatively large difference from the equipment we made in the past. In the past, X-Ray was more of an inspection equipment in the production and manufacturing process, with the core goal of finding defects. Besides production inspection, this sub-micron product can also enter scientific research institutes and laboratories to become a measurement equipment. Inspection and measurement are actually not exactly the same. Production inspection is more about judging whether there are defects; in the laboratory, besides inspection, attention must also be paid to exactly how long, how wide, and what the dimensions of the structure are. This places higher demands on precision, reliability, and consistency.
36Kr: Which part of Hik Ruiying's existing product matrix does it fill?
Wang Wei: Mainly two directions: one is semiconductors, especially advanced packaging; the other is laboratory-level failure analysis and measurement. This means our products are extending from primarily serving production inspection to higher-precision measurement. There are actually not many players globally who can truly enter this field; you can count them on two hands. There are only a very few domestic manufacturers capable of moving in this direction, and the entire industry is currently still in a relatively early stage.
36Kr: Benchmarking against overseas leading manufacturers, what level is this product at now?
Wang Wei: Overall, in the field of high-end X-Ray equipment, China has mostly followed overseas development in the past. But with the emergence of such high-precision equipment, the performance has basically reached the global first tier. In some local scenarios, we can also take the lead. For example, in complex defect detection, we can equip stronger AI models; in situations with a poor signal-to-noise ratio, we can also enhance through image algorithms to reveal defects that were originally difficult to see.
36Kr: From entering industrial inspection in 2021 to achieving sub-micron levels today, what is Hik Ruiying's technological evolution path?
Wang Wei: At the very beginning, we made security inspection equipment, which was basically millimeter-level; later we entered industrial inspection, reaching the micrometer level; now with the sub-micron product, we have taken another step forward. The core still revolves around two things: one is physical perception capability, seeking higher precision and stronger penetration; the other is inspection capability, seeking higher detection rates and fewer false alarms.
AI algorithms are also evolving simultaneously. From the very beginning of making inspection equipment, we actually rarely used traditional pattern recognition; the first generation of products already used deep learning. In recent years, we have gradually moved from small models to large models. The general direction has not changed, but the specific pace will change according to user needs. We do not have a particularly strong mandate to do a certain industry first and another later. It is more about the collision between X-Ray and AI integration and the real pain points of customers. Wherever the value is greater and the pain point is more urgent, we may prioritize working with users to build that scenario.
From a technical dimension, pursuing higher precision and stronger penetration in physical capability has no visible end. AI is the same; from pattern recognition to small models and then to large models, it may further move towards autonomous evolution in the future, making the equipment increasingly user-friendly with less human intervention. It is somewhat like the world record in sports. The record is there, but eventually, someone will push it forward a bit more. We do the same when making equipment; we must constantly challenge this limit.
II. The Game Changer: AI
36Kr: From the perspective of AI and X-Ray integration, what is the place where AI truly changes X-Ray?
Wang Wei: If we look back, X-Ray is a very mature technology. There have been no particularly major breakthroughs in core technology in the past ten years, mostly small iterations. So when you look back, you feel this is a traditional industrial equipment industry. Many companies have been doing it for ten or twenty years, possessing deep material experience, mechanical experience, and engineering accumulation. But if you look to the future, you will find this industry is actually particularly immature. The core is the integration of AI and X-Ray. In the past, X-Ray could capture internal images, but many things could not be automatically understood after being captured, so it could only be used as laboratory or spot-check equipment. Once AI can automatically understand these images, it has the opportunity to shift from "spot checks" to "100% full inspection."
Shifting from "spot checks" to "100% full inspection" directly determines whether this technology can achieve scale. In the past, the X-Ray applications that truly formed relatively large industries in industrial settings were actually just a few, such as food foreign object detection, PCBA solder joint void detection, and industrial flaw detection. Why can these scenarios achieve scale? Because they realized automatic judgment relatively early. Without automatic detection, a factory might only need one or two devices, placed in a laboratory for failure analysis and spot checks. But once full inspection is possible, a production line might need one device, and the equipment quantity will be completely different.
So I have always felt that X-Ray itself is a general capability. It can penetrate and can image. Theoretically, many products have a need to see internal defects. The fact that many scenarios did not use it in the past does not mean there was no demand; it is more that the automated inspection capability and the return on investment did not cross the threshold.
36Kr: Is it like industrial cameras ten years ago?
Wang Wei: Yes. I often look back at the development of industrial cameras from 2015 to 2020. Before 2015, it was actually hard for everyone to imagine that industrial cameras would be used so much on production lines. But later, with the decrease in costs and the improvement of AI and automated inspection capabilities, industrial cameras became a ubiquitous perception capability on production lines. X-Ray may be undergoing a similar process. Industrial cameras mainly look at the surface, while X-Ray looks inside. Many products have internal defects; it is just that in the past, the equipment was too expensive, or automatic detection could not be realized. Now, costs are decreasing, and capabilities are improving. These two bottlenecks are being opened simultaneously.
36Kr: Are there any typical cases?
Wang Wei: The foldable phone hinge is a very typical example. A hinge may contain hundreds of components. How was it inspected in the past? Many still relied on manual inspection with a magnifying glass. But a magnifying glass can only see the surface, and for such complex devices, humans will definitely have missed inspections after looking for a long time.
We started doing automatic X-Ray inspection of hinges a few years ago. At that time, the hinge was a new process facing pain points such as the outflow of various defects like foreign objects, cracks, voids, and missing assemblies. Through 100% AI X-ray inspection, the defect rate dropped by an order of magnitude. This is X-Ray combined with AI, turning things that were previously difficult to achieve full inspection into a "production process." There is also complete laptop inspection. The complexity is higher than that of hinges, but the essence is the same: whether screws are missing, whether there are extra ones, whether there are foreign objects—these are all things X-Ray is very suitable for.
36Kr: Is food inspection based on a similar logic?
Wang Wei: Yes, for example, there cannot be residual bones in chicken breast or chicken leg meat. After we entered this scenario a few years ago, during customer trials, we were already able to detect bones as thin as about 1 millimeter. Customers were also shocked at the time because this field was basically dominated by foreign players in the past. On this basis, we utilized the energy spectrum characteristics of X-Ray, combined with AI, to conduct fat-rate detection. For example, the ratio of fat to lean meat in pork is different, which affects taste and price. In the past, only a very few companies globally could do this, and the equipment was very expensive. Now that we have developed it, it has the opportunity to gradually become a more conventional capability.
I often say a sentence: When new technology collides with traditional technology, many sparks will fly. What we need to do is catch these sparks.
36Kr: Everyone is talking about large models now. What is the core value of large models truly landing in industrial X-Ray?
Wang Wei: A very important point in industrial AI is practical effectiveness. There are many open-source models today, and at first glance, people feel AI is already very mature. But when you truly return to the industrial site, you will find there are not many doing it well. Why? Because in many cases, industrial scenarios do not give you that much time. When a new model comes out, it might only take a month or two from sampling to mass production. Can you create a brand-new category within this time and meet customer requirements? This threshold is very high. If you give a company one or two years, many companies might be able to do it. But industrial customers will not wait for you for a year.
36Kr: Is small-sample learning very important in this?
Wang Wei: Extremely important. Using traditional methods to create a hinge inspection model might require thousands or even tens of thousands of images, followed by annotation and training. Based on the foundation of Hikvision's Guanlan large model, a brand-new category might only need one or two hundred, or a few hundred images to achieve good results. If this category has been done before, for example, if a type of hinge was done before and now a new model is changed, it might only take about ten images to migrate relatively quickly.
If you compare the model to a child, others might have to teach from scratch, but our child has already read up to elementary or even middle school. Teaching professional knowledge will naturally be faster.
36Kr: What is the actual value of this capability for, say, multi-variety, small-batch manufacturing?
Wang Wei: First is time. Going from 10,000 images to 200, the data preparation process alone might be dozens of times faster.
Second is precision and false alarms. Industrial inspection ultimately depends on results. Many products now require 3 nines (99.9%) or 4 nines, and customers will continue to demand you move towards 5 nines.
Third is generalization ability. Originally, one model was trained for one product. Now, if the foundation is strong enough, after I have done one category, the data needed to do similar categories will be less and less. So what users feel is that the equipment is becoming increasingly "user-friendly."
36Kr: Besides defect recognition, what other links has AI entered now?
Wang Wei: Actually, there are many. For example, in the past, a device contained a large number of image parameters. How to adjust the gamma value, how to adjust the grayscale—these required very professional personnel. Now with AI, we can achieve one-click parameter adjustment. The equipment automatically adjusts to relatively reasonable parameters based on the image, and the user operation threshold will decrease.
Image enhancement is the same. An image is essentially signal and noise. Sometimes defects are unclear because the signal is suppressed by noise. AI can learn noise features, strip away the noise, and let the real signal emerge. These capabilities have been implemented for three or four years.
Looking further to the future, I think "autonomous evolution" is worth paying attention to. Whether the equipment can become increasingly user-friendly and smarter with less human intervention—this may be where changes continue to happen in the next few years.
36Kr: Electronics, food, and automotive components differ greatly. How does the Hikvision Guanlan large model adapt across industries?
Wang Wei: We have a relatively simple phrase ourselves: "If it can be seen, it can be inspected." As long as the defect has visible features in the X-Ray image, AI has the opportunity to solve it. On the surface, one is an electronic device, one is food, and one is an automotive wheel hub—they are completely different. But for the model, they are ultimately X-Ray pictures. Normal images may hide different defect features, which we quickly identify and iterate through the underlying large model. So although the industries look different, the bottom line ultimately returns to "AI + X-Ray."
36Kr: AXI is known in the industry as Automated X-Ray Inspection. Now Hik Ruiying proposes AI X-Ray Inspection. What does this change mean?
Wang Wei: "Automated" is certainly not wrong, but we hope to emphasize the "AI" aspect a step further. Many traditional devices still stay at pattern recognition, which is only suitable for relatively simple scenarios. From the very first device, Hik Ruiying implanted deep learning inside. Originally it was small models, and now it is gradually shifting to large models.
The first change is a significant improvement in detection capability under complex backgrounds; the second change is more important, which is that AI begins to possess a certain level of understanding. In the past, inspection equipment only told you "there is a defect here." In the future, it can further tell you what kind of defect it is. Going further, it can also inversely analyze the front-end process and judge which parameter might have gone wrong. This forms a closed loop from inspection: discovering defects in the back-end, and then inversely optimizing the front-end process. The deeper AI penetrates, the higher the value of this equipment will be.
36Kr: So will future competition lean more towards AI competition?
Wang Wei: Both hardware and AI are important. The physical capability of X-Ray itself still needs to continue breaking through—higher precision, stronger penetration; this is always important. But looking at the opportunities at this current stage, the speed of change in the integration of AI and X-Ray is faster, and there are more scenarios that have not been implemented, so the opportunities may be greater.
I have always had a judgment: the future will definitely be intelligent equipment. It will not just be traditional equipment; in the future, it can even be understood as a kind of "equipment robot," possessing more and more capabilities for automatic judgment, automatic execution, and self-adjustment. In fact, over the past five or six years, our market performance has also been gradually verifying this judgment.
III. "Looking Back, It's a Red Ocean; Looking Forward, It's a Blue Ocean"
36Kr: Many people's first impression of Hikvision is still security. What role does Hik Ruiying play among the eight innovative businesses?
Wang Wei: Hikvision is now more than just security. Over a decade ago, we were already considering future directions. Later, we gradually extended the perception capabilities in security to smart homes, automotive electronics, industrial production, and other fields.
Another dimension is that the perception methods themselves are expanding. In the past, it was mainly visible light; now it extends to infrared thermal imaging, millimeter-wave, X-ray, sound waves, and other directions. Hik Ruiying carries Hikvision's layout in the fields of X-Ray and non-visible light detection. Simply put, it is looking inside through the surface.
36Kr: If Hik Ruiying is placed within Hikvision's entire industrial business, is it a microcosm of Hikvision's deep cultivation in the industrial field?
Wang Wei: It can be understood this way. Now Hikvision provides products and solutions in various fields such as industrial quality inspection, equipment monitoring, production operations, and safety production. For example, in industrial inspection, besides X-ray products, there are also ultrasonic non-destructive testing series products, super-depth-of-field 3D digital microscopes, mass spectrometers, chromatographs, and other equipment. These products have different technological routes, but the industrial trends they face are very similar; they are all integrating with AI to find high-value applications. In industrial inspection, we are actually slowly building a more comprehensive perception. For example, for a chip, ultrasound is more suitable for looking at planar layer cracks; a super-depth-of-field microscope looks at microscopic surface defects; if it is finally found to be a material problem, chromatographs and mass spectrometers might be used for quantitative and qualitative analysis.
Any technology has its strengths and weaknesses. Hikvision formed a "multi-dimensional perception" technological route since the security era. The toolbox has different perception capabilities, which are then combined according to user scenarios.
36Kr: How does Hikvision's own manufacturing system help Hik Ruiying's products?
Wang Wei: This help is very important. Many of our products will be used in batches in Hikvision's factories first. Some applications might be unexpected to the outside world. For example, we use them to check if there are missing accessories in the camera packaging. Because X-Ray is essentially penetration. No matter how well you design a product on paper, it must go to the battlefield. Only in the real manufacturing scenario can you know where it is good and where it is not. We are quite lucky to be able to rely on Hikvision's own manufacturing scenarios for full verification first, and then deliver mature solutions to external customers.
36Kr: As a relatively young industrial inspection brand, did you encounter customer misunderstanding in the early stages?
Wang Wei: Definitely, and it's quite interesting. When we first actively pushed it outwards, the effect was actually not particularly good. Later, many customers saw our capabilities at exhibitions or through other channels and would proactively come to us. Once we hit real problems, the stickiness was actually very strong. Why? Because we solved problems that could not be solved in the past. So our business logic has not been entirely arranged by industry. I do not first stipulate doing Industry A this year and Industry B next year. The core is where value can be generated after AI and X-Ray integration. As more and more such cases emerge, the industry's recognition of Hik Ruiying's capabilities is slowly established. In the final analysis, it is about making good products, solving problems, and creating value for customers.
36Kr: Are there any particularly "atypical" industrial scenarios?
Wang Wei: Many. For example, eyebrow pencils. Some eyebrow pencils have springs inside. If the spring is not installed well, it will get stuck. In the past, it was hard to see inside, so we used X-Ray to judge if the spring was installed properly. There are also automotive temperature sensors and motors, which were not typical production inspection scenarios for X-Ray in the past. Such examples are increasing. Because customers' quality control requirements are improving, and equipment costs are decreasing. Once costs come down, many industries that would never have considered X-Ray in the past can now make the math work.
36Kr: How do you view the current industrial stage of industrial X-Ray?
Wang Wei: X-Ray security inspection has passed the initial stage and entered the early-to-mid stage. I think industrial X-Ray has just begun and is still in the early stage.
Therefore, looking back, it's a red ocean; looking forward, it's a blue ocean. Looking back, there are many companies that have been doing X-Ray for ten or twenty years, and traditional scenarios are also very mature. But looking to the future, AI is opening up new scenarios, costs are decreasing, so there are still many things that have not been created.
36Kr: Will these new inspection capabilities eventually gradually become standard processes in manufacturing?
Wang Wei: I think they will. Today, a customer might take the lead in adopting a certain inspection technology, and their quality will suddenly be much better than their peers. When terminal customers see this, they will put forward the same requirements for other suppliers.
Today's highest requirements are very likely to be tomorrow's lowest standards. So many technologies that look customized at the beginning may become a "small standard product" or even a standard process in the industry after three or five years.
This process is dynamic. Our technology is advancing, industry requirements are improving, and standards are constantly changing. There is inherently strong momentum in this.
36Kr: Compared with overseas traditional X-Ray manufacturers, where might Chinese enterprises gain opportunities in the future?
Wang Wei: If we only look at the top-tier X-Ray physical capability, we are still catching up overall. But when it comes to the equipment level, AI will definitely become a very important competitiveness. From my own contact, I have a relatively obvious feeling: many excellent overseas X-Ray companies actually do not integrate AI with specific scenarios very much.
On the one hand, it is a capability issue. Hikvision itself has a trillion-level business volume, with many perception businesses supporting the Guanlan large model together. This kind of foundation is not something every X-Ray company can possess.
On the other hand, it is the business model. Industrial AI is often essentially customization and service. A scenario might end up selling only a few or a dozen devices. We are willing to do such things, but many overseas companies may not be willing to.
China has a complete manufacturing supply chain and a large number of real scenarios. The distance between technology and the market is relatively close, so many new capabilities can quickly penetrate into segmented industries.
36Kr: What is the current progress of Hik Ruiying in several industrial business segments?
Wang Wei: Several fields are developing well now. Electronic manufacturing, food, and industrial-related inspection all have very broad scenarios that have not been covered by X-Ray. After several years of development, we can basically rank in the top three in multiple segmented equipment fields domestically. It is hard for me to say which one will definitely have greater opportunities in the future, because looking at it now, every direction has a lot of space.
36Kr: Will the overseas market become the next focus?
Wang Wei: Yes. China's high-end equipment capabilities will continue to move forward with the development of the entire manufacturing system. Hik Ruiying is not a single enterprise moving forward alone; behind it, it actually undertakes the improvement of the capabilities of China's entire X-Ray supply chain and manufacturing system. I believe that Chinese equipment will also, like shoes and clothing, white goods, mobile phones in the past, and including new energy vehicles now, move towards the global market more and more. For us, the overseas market is definitely very important.
36Kr: Is there a relatively clear global goal?
Wang Wei: We have a relatively clear vision internally for the short to medium term: to be the best AIX globally. But if we must give a very quantified definition to "the best," I think it is actually difficult. For me, it ultimately comes down to users finding it user-friendly. We believe that standing on Hikvision's platform, combined with our own efforts and the overall improvement of China's manufacturing capabilities, we have the opportunity to do this.
36Kr: If ten or twenty years from now, people mention X-Ray inspection again, what do you hope they think of first?
Wang Wei: Think of Hik Ruiying. To be more specific, I hope everyone thinks of: Hik Ruiying, at this stage, led this trend of X-Ray intelligence. We were founded in 2020 and have walked for a few years now. Looking back, around 2020 was actually a turning point for smart security inspection. At that time, on the one hand, we were promoting the integration of X-Ray and AI; on the other hand, the trend of this era was also成就 (achieving/shaping) us.
Just as the people of Zhejiang say "tide riders"—at this time point, you must stand on the crest of the tide. If the crest of the tide is right, the future is yours. At least you have the opportunity.
Sources
- 36Kr (2026-10-13) - Original link: Interview with Wang Wei of Hik Ruiying: Century-Old X-Ray Technology is Being Redefined by AI - https://m.36kr.com/p/4022541911412609