The science of sweetness: the next generation of blueberry intelligence

For decades, automated fruit grading has been fundamentally visual. Color, size, shape, and external defects have provided the basis for increasingly sophisticated sorting systems. The next frontier is different: understanding what lies beyond appearance and translating internal quality into actionable data at commercial scale.

GP Graders is advancing this frontier with gpVision™ and radiai™, its next-generation fruit intelligence platform combining sophisticated multi-camera imaging with deep-learning artificial intelligence to predict blueberry sweetness non-destructively, in real time and at production-line speeds.

The technology integrates monochrome, color, and near-infrared (NIR) imaging to create a rich optical profile of each blueberry. radiai™ interprets this multidimensional information to predict sweetness, transforming complex spectral and visual signals into a practical quality metric.

GP Graders adds smarts to blueberry sorting 

In recent blueberry testing, the system achieved a mean absolute error of 1.15 degrees Brix and reported 91.71 percent average accuracy. Results included 53.2 percent of samples within ±1.0 degrees Brix, 70.1 percent within ±1.5 degrees Brix, and 76.6 percent within ±2.0 degrees Brix.

These results point towards a significant evolution in automated grading: from identifying what a blueberry looks like to predicting what it is likely to taste like.

This distinction is central to GP Graders’ technology leadership. Rather than treating sweetness as a laboratory attribute measured separately from the grading process, its approach brings advanced imaging and AI together to make sweetness prediction part of high-speed fruit intelligence.

For growers, packers and exporters, the implications are substantial. The ability to generate non-destructive sweetness data at scale opens new possibilities for quality segmentation, pack-out optimization, and greater consistency across global supply chains.

The future of fruit grading will not be defined solely by better cameras. It will be defined by better intelligence. 

*All images by GP Graders.

Related stories

The best innovations in eco-friendly materials for fresh produce

Copefrut and Sinclair strengthen their commitment to sustainable fruit packaging