The OASYS project focuses on optoelectronic sensors for application-oriented systems. In subproject A1, a team of scientists and industry experts is developing an ultra-compact, energy-efficient hyperspectral camera that uses artificial intelligence to perform complex materials and quality analysis in real time. An integrated spectrometer records spectral characteristics and reveals chemical properties invisible to the human eye. This allows food defects and the composition of fibers and plastics to be identified quickly and accurately.
New hyperspectral cameras offer a wide range of applications in industrial and agricultural processes. Its innovative approach combines traditional 2D imaging with artificial intelligence and spectral analysis. A standard 2D camera first captures a high-resolution image of the target object. Artificial intelligence then analyzes the image in real-time and automatically identifies areas of interest. An integrated spectrometer then performs spectral analysis only at these selected locations to determine the chemical composition. This intelligent approach increases the efficiency of hyperspectral measurements. Rather than acquiring spectral data for the entire image (a computationally intensive process), the system analyzes only relevant measurement points. This targeted approach reduces data volume, energy consumption, and processing time.
The information thus obtained supports, for example, the reliable classification of textiles and plastics. It also increases confidence in identifying counterfeit products. Additionally, it improves quality control in food processing by detecting pressure marks and defects, and enables accurate assessment of plant condition and nutrient requirements in agriculture. Automated assessments enable faster and more reliable decision making. At the same time, processes become more sustainable and economic resources are used more efficiently.
“With the OASYS A1’s compact hyperspectral camera, we are developing a technology that can be used directly in production lines, sorting facilities or in analytical processes in the field,” explains Heinrich Engelke, project manager at the Fraunhofer Institute for Photonic Microsystems IPMS. “The combination of miniaturization, energy efficiency and artificial intelligence opens up completely new applications, while also making important contributions to resource savings and process reliability.”
The components developed in this project will form the basis of future sensor systems that can improve the industrial, recycling, agricultural and food sectors.
