Utilizing Machine Vision to Enhance Unattended Sample Preparation through Automatic Error Detection and Mitigation

Automation & Innovation for Sample Preparation
Poster Presentation

Prepared by J. Nichol, I. Wan
PromoChrom Technologies, 13351 Commerce Parkway, Richmond, British Columbia, V7E 1Z7, Canada


Contact Information: [email protected]; 18337724766

ABSTRACT

Sample preparation, such as Solid Phase Extraction (SPE), plays a critical role in analytical instrumentation. Due to the tedious and time-consuming nature of these processes, laboratories increasingly rely on automated systems to improve efficiency. While these systems aim to free up personnel time, they are designed to perform a very specific and repeatable set of processes. Nuances arising from increasingly challenging samples, operator error and machine wear can lead to unexpected disruptions that require human intervention. Without real-time monitoring, labs may only detect issues after failed runs or sample loss, leading to costly rework.

This presentation explores how Machine Vision can enhance the reliability of unattended liquid handling by enabling real-time error detection and mitigation. Using deep learning, the system is trained to monitor critical liquid handling components, recognizing common operational patterns and identifying anomalies similarly to the human eye. By leveraging image recognition and an expanded training dataset that includes edge cases, the system can automatically detect unforeseen issues. This monitoring technique enables equipment software to apply corrective actions and self-check, ensuring sample preparation continues with minimal disruption. With these implementations, high throughput and even overnight sample preparation can be achieved with greater confidence.

As laboratories move toward further automation to reduce manual oversight and increase throughput, Machine Vision serves as a critical safeguard — providing the necessary "set of eyes" to ensure smooth and reliable operation.