The problem we aimed to tackle is the accurate measurement of liquid height in a transparent glass using only an image.
Traditional measurement methods rely on direct physical contact, but in scenarios where automation, remote estimation, or digital analysis is required, a computer vision-based approach becomes essential.
This study focuses on developing a non-intrusive method to estimate the liquid level in a glass, which can be applicable in various domains such as:
- Food and Beverage Industry: Automated quality control in beverage filling lines.
- Healthcare and Nutrition: Monitoring fluid intake for medical applications.
- Smart Kitchens: Estimating liquid levels in smart cooking and IoT-enabled devices.
To address this challenge, we began by creating a dataset of proper images and developed an intuitive initial approach, which we progressively refined by making adjustments based on the issues we encountered.