Protostar Labs proposes a research project in collaboration with the University of Zagreb Faculty of Electrical Engineering and Computing (FER) to explore the possibilities of designing an innovative, flexible, and cost-effective multispectral camera. The multispectral camera will combine an affordable high-resolution RGB camera with specially designed optical filters tailored for analyzing and solving specific problems. Additionally, the goal is to integrate computer vision and AI algorithms to enable automatic material detection and analysis.

Industry challenges can be addressed by applying industrial cameras and AI-based computer vision software. Although such systems are becoming increasingly popular in factories worldwide, their widespread adoption is limited by high costs. Multispectral cameras, for example, can cost over €25,000. In large manufacturing plants where multiple cameras are needed, these costs present a significant barrier.

Initially, the prototype multispectral camera will be tested to address two specific problems: detecting contamination in food production and remote soil monitoring for agricultural planning. If the research yields positive results, it could completely transform the approach to and accessibility of multispectral analysis for solving problems in various industrial sectors.

Furthermore, the partners will explore the latest AI algorithms, including deep learning and neural networks, for processing and analyzing collected data. These technologies will enable more precise object classification and recognition, as well as real-time decision-making.

With this project, Protostar Labs is advancing the industrial vision systems sector, aiming to provide accessible and high-quality technological tools. The research results will have a revolutionary impact on various industries, from the food sector to space exploration, contributing to solving key societal challenges and paving the way for new scientific discoveries and technological innovations.

Protostar’s team also specializes in developing remote sensing solutions using multi-/hyperspectral cameras on drones and satellites. Such solutions enable large-scale Earth monitoring and provide crucial data for analyzing and predicting environmental changes and phenomena. Protostar Labs is already collaborating with ESA and FER in developing AI technologies for satellite applications in low Earth orbit.