In the pharmaceutical industry, the quality detection of vials and ampoules is of great importance as it is related to medication safety. However, traditional detection methods face numerous difficulties when dealing with transparent container detection. WeLinkirt's DaoAI AI-AOI high-speed light inspection solution brings new hope for solving this problem.
Industry Background and User Scenarios: In the pharmaceutical industry, vials and ampoules are the most common packaging containers for injections. Their quality is directly related to the safety and effectiveness of drugs. During the drug production process, the filled and sealed vials and ampoules need to undergo strict inspections to ensure that the liquid level is within the specified range, the sealing is intact, and there are no transparent foreign objects or glass chips in the bottles. These inspection links are key steps to ensure medication safety. For pharmaceutical production enterprises, an efficient and accurate inspection solution is essential.
Pain Points: Why is it Difficult?
From a quantitative perspective, there are numerous difficulties in the inspection of vials and ampoules. First of all, in terms of refracted light interference, the similarity between glass chips and the refracted light from the bottle wall is over 90%, which makes it very difficult for traditional visual inspection equipment to accurately distinguish them. Secondly, in the identification of minor defects, the micro-cracks in the fused ends of ampoules may be between 10-50 microns in size, and the slight lifting height of the rubber stoppers of vials may only be 20-100 microns. Such minor defects are almost imperceptible to the naked eye, and it is also very difficult for traditional inspection equipment to accurately identify them. Moreover, the rhythm of the high-speed filling and sealing line is extremely fast. Usually, the inspection time for each bottle is only 0.5-1 second, which poses extremely high requirements for the response speed and processing ability of the inspection equipment.
The root cause of these inspection difficulties lies in the characteristics of the transparent system itself. The clear liquid makes it easy to confuse glass chips with the refracted light from the bottle wall, just as it is difficult to find a transparent object in clear water. The micro-cracks in the fused ends of ampoules and the slight lifting of the rubber stoppers of vials are very subtle and difficult to detect in a fast-paced production line. Moreover, on a production line with a very high pass rate, defective samples such as glass chips and micro-cracks in sealing are extremely rare. It is very difficult for traditional inspection methods to accurately train models with limited samples, which greatly reduces the accuracy and reliability of the inspection.
Technical Principle
The DaoAI AI-AOI high-speed light inspection solution uses advanced algorithms, imaging, and hardware mechanisms. In terms of algorithms, deep learning algorithms are used to judge the liquid level, sealing, and foreign objects. Deep learning algorithms have powerful feature extraction capabilities and can accurately identify whether the liquid level is within the tolerance band, the sealing integrity of the stopper/fusion sealing, and glass chips and foreign objects in the transparent liquid from complex images. Through a large number of sample trainings, the algorithm can be continuously optimized to improve the accuracy of the inspection.
In terms of imaging, the solution uses high-resolution imaging equipment, which can take a single image at the post-filling and sealing station. High - resolution images can clearly capture the details of the bottle and provide accurate data for subsequent analysis. At the same time, through special lighting technology, the reflective interference on the liquid surface is reduced, making the image clearer. Compared with traditional methods, traditional visual inspection methods can usually only detect a single indicator and are easily affected by refraction and reflection, resulting in low accuracy and efficiency of the inspection. The DaoAI AI-AOI high-speed light inspection solution can take a single image and conduct simultaneous judgment, greatly improving the inspection efficiency and accuracy.
Typical Application Scenarios
- Liquid Level Detection: During the filling process, the accuracy of the liquid level is crucial. The DaoAI AI-AOI high-speed light inspection solution uses deep learning algorithms to accurately judge the liquid level. The difficulty lies in that the refraction and reflection of the transparent liquid will affect the accurate identification of the liquid level. However, the solution effectively solves this problem through special image processing technology and algorithm optimization.
- Sealing Detection: This includes the fusion sealing of ampoules and the stopper sealing of vials. The micro-cracks in the fused ends of ampoules and the slight lifting of the rubber stoppers of vials are the difficulties in the inspection. The solution uses high-resolution imaging and deep learning algorithms to accurately identify these minor defects. For example, it can also achieve high-sensitivity detection for micro-cracks of 10-50 microns in the fused ends of ampoules.
- Glass Chip Detection: It is very difficult to find glass chips in the transparent liquid. The solution is specifically designed for high-sensitivity identification of glass chips and foreign objects. Through special lighting and image processing technology, the contrast between the glass chips and the background is enhanced, improving the accuracy of the detection.
- Foreign Object Detection: In addition to glass chips, other transparent foreign objects may also be mixed in the bottle. The solution analyzes the shape, size, and position of foreign objects through deep learning algorithms to achieve effective detection of various transparent foreign objects.
Implementation Case
There is a medium-sized pharmaceutical production enterprise that mainly produces injections. Before introducing the DaoAI AI-AOI high-speed light inspection solution, the enterprise used traditional visual inspection equipment, which had low inspection efficiency. The detection rate of transparent foreign objects and glass chips was only 80%, and the miss-detection rate of sealing defects and liquid level deviations was as high as 15%. Moreover, due to the slow inspection speed, it affected the overall rhythm of the production line, resulting in low production efficiency.
Transparency never means simplicity. Only by overcoming refraction, reflection, and rare defects can we truly guard the final pass of injections.
During the implementation process of the DaoAI AI-AOI high-speed light inspection solution, the technical team of WeLinkirt conducted a detailed on - site investigation and equipment debugging. According to the characteristics of the enterprise's production line and inspection requirements, the solution was optimized in a personalized way. After a period of debugging and optimization, the solution was successfully launched. After the launch, the on - line full inspection of the liquid level, sealing, and glass chips in the filling and sealing of vials and ampoules was realized in the factory. The detection rate of transparent foreign objects and glass chips increased to over 98%, and the miss-detection rate of sealing defects and liquid level deviations decreased to below 2%. It ran stably under the high-speed rhythm and passed the GMP verification audit completely.
WeLinkirt's Solution and Product
WeLinkirt's DaoAI AI-AOI high-speed light inspection solution is a comprehensive inspection solution. The solution integrates advanced deep learning algorithms, high-resolution imaging equipment, and special lighting technology, and can simultaneously complete the inspection of liquid level, sealing, and foreign objects on the basis of a single image. At the same time, for rare defects such as glass chips and micro-cracks in sealing, the APDT few-shot learning method is used to quickly build models and go online. The whole-process data can be audited, which meets the GMP requirements and provides pharmaceutical production enterprises with an efficient, accurate, and reliable inspection method.
Quantitative Results: By implementing the DaoAI AI-AOI high-speed light inspection solution, the enterprise has achieved significant improvements in multiple indicators. The detection rate of transparent foreign objects and glass chips has increased from 80% to over 98%, an increase of 18 percentage points; the miss-detection rate of sealing defects and liquid level deviations has decreased from 15% to below 2%, a decrease of 13 percentage points; at the same time, the inspection efficiency has been greatly improved. Under the rhythm of the high-speed filling and sealing line, the inspection tasks can be stably completed, ensuring the efficient operation of the production line.
FAQ
What challenges does the inspection of vials and ampoules face?
The inspection of vials and ampoules faces great challenges. In the transparent system, it is difficult to distinguish glass chips from the refracted light of the bottle wall, with a similarity of over 90%. The micro-cracks in ampoules and the slight lifting of rubber stoppers in vials are difficult to identify, with sizes ranging from 10-100 microns. The inspection time on the high-speed filling and sealing line is only 0.5-1 second, and defective samples are scarce. WeLinkirt can handle these challenges.
How does the DaoAI AI-AOI high-speed light inspection solution address the inspection challenges?
This solution takes a single image and conducts simultaneous judgment. It uses deep learning to determine the liquid level, detect the sealing integrity, and identify foreign objects with high sensitivity. The APDT few-shot learning method is used to handle rare defects, and the whole-process data meets the GMP requirements, effectively improving the accuracy and efficiency of the inspection.
What effects does the solution have after its launch?
After the launch of the solution, the on - line full inspection of the liquid level, sealing, and glass chips in the filling and sealing of vials and ampoules can be realized. The detection rate of transparent foreign objects and glass chips has increased to over 98%, and the miss-detection rate of sealing and liquid level has decreased to below 2%. It runs stably and passes the GMP verification audit.
This article was generated by AI. Customer cases are simulated scenarios based on real product capabilities and figures are illustrative; see product pages for official benchmarks.