Methods and Apparatuses for Detecting Anomalies When Filling a Container With Fluid Watching Each Bottle Fill and Catching Faults in Real Time

This patent describes a machine vision approach that watches video of a container being filled with liquid and detects problems as they happen. If something goes wrong during filling, the system can raise an alert or adjust the filling machine on the spot, which is especially valuable in pharmaceutical and other high-stakes production.

The Problem

Filling containers with liquid at high speed is a routine but unforgiving step in manufacturing, particularly in aseptic processing of medicines where every unit must meet strict standards. Faults such as overfilling, underfilling, splashing, foaming, bubbles or foreign matter can compromise a product. Traditional checks often happen after filling, which means defective units are only caught once they are already made, wasting material and time. Catching an anomaly during the fill, while there is still a chance to react, is much harder because it requires interpreting fast-changing images in real time.

What This Invention Does

The patent provides computerised methods and apparatuses that use one or more images from a video of the filling process to detect anomalies within a container as it is filled. The techniques are designed to work in real time, analysing the fill as it happens rather than afterward. When at least one anomaly is detected, the system can output an alert or alter one or more parameters of the filling process that the machine is configured to perform. In practice this means the line can flag a suspect unit immediately, or even adjust filling settings automatically to correct course, reducing waste and protecting quality.

Key Features

• Video-based inspection. The system analyses images from a video of the container being filled.
• Real-time detection. Anomalies are identified during the fill, not only after it is complete.
• Automatic response. On detecting a fault, the system can raise an alert or change filling parameters.
• Process feedback. The ability to alter machine settings supports correction on the line.
• Suited to demanding production. The approach fits high-standard environments such as pharmaceutical filling.

Who Is Behind It

The applicant is Eli Lilly and Company, a major global pharmaceutical manufacturer. The named inventors are Hetalkumar Meghrajbhai Desai, Matthew Glenn Kawiecki, Rakesh Sharma, Asit Tiwari and Jordan Matthew Venderley. The application is a divisional of an earlier filing.

Why It Matters

In regulated manufacturing, quality control that catches defects earlier saves both money and risk, and real-time vision systems are becoming central to modern production lines. Detecting fill anomalies as they occur reduces scrap, supports compliance, and helps keep high-speed lines running reliably. Protecting the method in Australia supports Lilly’s manufacturing and supply interests in the region.

Related Concepts

• Machine vision – the automated imaging technology behind the inspection.
• Anomaly detection – the analytical task of spotting faults in data.
• Aseptic processing – the sterile filling context where this matters most.
• Quality control – the broader manufacturing goal the system serves.
• Computer vision – the field of teaching computers to interpret images.

---

AU 2026201904 was published in the Australian Official Journal of Patents on 2 April 2026 and is open for public inspection. Patent applications represent inventions that are sought to be protected and do not necessarily reflect commercially available products.