AI-Enabled Compact Hyperspectral Camera Opens New Paths for Industrial Analysis

The OASYS research project is advancing optoelectronic sensor technologies designed for application-specific systems. Within subproject A1, researchers and industry specialists are developing an ultra-compact, energy-efficient hyperspectral camera that integrates artificial intelligence to enable real-time material and quality analysis. By capturing spectral data beyond the visible range, the system identifies chemical properties that cannot be detected by the human eye, supporting faster and more precise detection of defects in food products and analysis of textile and plastic compositions.

The hyperspectral camera is designed for use across industrial and agricultural environments. Its technical concept combines conventional 2D imaging, artificial intelligence, and targeted spectral analysis. A standard 2D camera first records a high-resolution image of the object. Artificial intelligence then evaluates the image in real time to automatically identify areas of interest. Spectral analysis is subsequently performed only at these selected points using an integrated spectrometer to determine chemical composition.

This selective measurement strategy improves the efficiency of hyperspectral imaging by avoiding full-image spectral capture, which is typically data- and energy-intensive. By limiting analysis to relevant regions, the system significantly reduces data volume, energy consumption, and processing time while maintaining analytical accuracy.

The resulting data enables more reliable sorting of textiles and plastics, improved detection of counterfeit products, and enhanced quality control in food processing, including the identification of pressure marks and defects. In agricultural applications, the camera supports precise assessment of plant health and nutrient requirements. Automated analysis allows for quicker decision-making, increased process reliability, and more efficient use of resources.

According to Heinrich Engelke, Project Manager at the Fraunhofer Institute for Photonic Microsystems IPMS, the compact hyperspectral camera developed under OASYS A1 is intended for direct deployment in production lines, sorting systems, and field-based analytical applications. The combination of miniaturisation, energy efficiency, and artificial intelligence is expected to support new use cases while contributing to resource conservation and operational reliability. The project’s components are also intended to serve as a foundation for future sensor systems across industry, recycling, agriculture, and food processing.