Céline Portenier

Switzerland has one of the world’s densest networks of outdoor webcams, both publicly accessible and privately operated. Thousands of images are recorded every day. These images not only show beautiful landscapes, but also contain information on the environment and current weather conditions. Today, these webcams are typically checked manually, which is slow, subjective, and inefficient, leaving a large source of localized, real-time information unused.

WebcamInsight aims to bridge the gap between weather model output and real-world conditions. Using computer vision and geospatial modeling, WebcamInsight turns webcams into automated measurement systems and extracts information such as fog and low cloud layer boundaries, cloud cover, visibility, snow cover, and vegetation status.

Beyond monitoring, Céline and her team address a key limitation of weather forecasting. Fog and low stratus clouds remain difficult to predict, as numerical and AI-based models often lack information on near-surface conditions. They provide data for model validation and short-term forecast improvement, particularly for AI-based approaches that depend on realistic input data.

WebcamInsight is designed for users that rely on accurate, location-specific weather information. This includes aviation and rescue services, energy and grid operators, transport authorities, weather services, and tourism organizations. These sectors benefit from better situational monitoring, more reliable forecasts, and a stronger basis for timely and confident decision-making.

During the UZH Entrepreneur Fellowship, Céline and her team will integrate observations from satellites and ground-based measurements into their webcam analysis and investigate whether these combined observations can improve fog and low-cloud forecasts.

Affiliation:  Dr. Claudia Röösli
Start date: 1.12.25