The cherry sorting, selection, and packing sector is undergoing a radical transformation driven by artificial intelligence (AI). By optimizing sorting systems, enhancing quality control, and enabling predictive maintenance, AI is making packing lines faster, more precise, and cost-effective. This evolution significantly impacts operations handling large volumes quickly, where precision and efficiency are paramount.
Challenges and Goals
Cherries exhibit great variability in quality, color, shape, size, and stem length, posing a complex analysis challenge. Being small and delicate, cherries are prone to damage. Common defects include cracks, rot, insect damage, bruises, wind damage, sunburn, and issues like overly soft texture or missing stems. Special care is required to prevent cracking due to their semi-permeable skin.
Therefore, the goal in packing facilities is to maximize yield from high-quality fruit. Fortunately, the industry can now adopt a tailored solution: artificial intelligence.
Designed, Trained, and Ready to Excel
The term “artificial intelligence” was coined in the mid-1950s, but its widespread adoption faced earlier challenges. Now, AI is experiencing a surge in popularity, fueled by deep learning, which leverages vast datasets and modern computing power.
“This was the starting point for GP Graders’ radiai system. With our extensive cherries sorting data, we could let this new technology aid us,” says Stuart Payne, GP Graders’ director.
“For the past five years, GP Graders has meticulously indexed and annotated hundreds of thousands of images of cherry defects, building one of the industry’s most comprehensive datasets. This extensive database allows AI to detect and classify defects with remarkable accuracy. Unlike traditional rule-based sorting systems that rely on predefined parameters, radiai uses advanced convolutional neural networks (CNNs) to automatically identify and differentiate a wide range of defects without explicit programming. By continuously improving through machine learning, the system adapts to cherries’ natural variability, enhancing both precision and efficiency.
“The core of this technology is AI’s ability to associate subtle visual patterns with specific defect categories, mimicking a well-trained human sorter’s experience but at unprecedented speed and scale. For instance, unusual discoloration may indicate underlying bruises, while irregular surface textures could signal cracks or scar damage. Cherries with irregular shapes are instantly identified for reclassification, ensuring only top-quality fruit goes through sorting.
“Thanks to deep learning, GP Graders’ radiai continuously refines its defect detection capability, leading to more precise, consistent cherry sorting adaptable to high-speed packing operations’ changing challenges.”
Proven impact of AI
Operators using gpVision with radiai report nearly perfect detection of stems, regardless of how many appear in a single image. This advancement addresses a common challenge in cherry sorting: misidentifying the stems of neighboring cherries, leading to inaccurate size measurements and nonexistent defect detection.
By precisely isolating each cherry and its respective stem, radiai ensures accurate diameter calculations, resulting in better calibration and improved packaging efficiency without sacrificing sorting accuracy. Unlike traditional systems that struggle with accuracy at high volumes, radiai seamlessly manages increases, boosting both productivity and efficiency.
“A significant improvement in defect detection is due to radiai’s ability to differentiate stems from fruit imperfections. Previously, a brown part of a stem might be misclassified as a crack, unnecessarily rejecting perfectly good cherries. With AI refining the process, this issue has been significantly reduced, minimizing false positives and unnecessary waste.
“In the past, each misclassified cherry resulted in a direct financial loss by wrongly discarding high-quality fruit due to a false positive. By enhancing sorting precision, radiai ensures only truly defective fruit is discarded, maximizing packaging rates and producer revenue.”
One of the most critical aspects of cherry sorting is crack detection, as even minor surface damage can affect a cherry’s classification. However, not all cracks impair quality equally; location matters. The radiai system intelligently distinguishes cracks in various locations, allowing for more accurate classification. A small crack near the bottom, for instance, may permit classification as second class, while a side crack, which likely impacts fruit integrity, would lead to rejection. Additionally, the system assesses stem quality, identifying dehydration or browning levels, key indicators of freshness and shelf life.
“One of the most revolutionary advancements of gpVision with radiai is its ability to detect recent cracks, a very challenging defect to identify. These micro-tears in the cherry’s skin are often too subtle for standard sorting systems. However, thanks to access to a vast and evolving dataset, radiai has nearly eliminated this issue. By recognizing even the most imperceptible defects, the technology provides producers with greater assurance that only impeccable, high-quality cherries make it into boxes and packages, ensuring customer satisfaction and avoiding costly returns.”
Reducing operator workload
Now that manual sorting is nearly obsolete, there’s a chance to enhance machine operators’ experiences and reduce the training time for new users.
With radiai technology at the core of gpVision, the operator experience has transformed completely. What was once a complex process requiring meticulous adjustments based on multiple data points has become an intuitive and user-friendly interface. Instead of manually adjusting complicated settings in complex spreadsheets, operators now have access to streamlined controls, featuring easy-to-use sliders and visually intuitive configurations that simplify the entire sorting process.
“Compared to other systems on the market, gpVision stands out for its superior ease of use. Achieving exceptional sorting accuracy has never been easier; operators simply review the data, make minor adjustments using intuitive sliders, and let the AI-driven system handle the rest. This mode of interaction not only reduces the learning curve for new users but also enhances efficiency, allowing packing facilities to optimize sorting accuracy with minimal manual intervention. The result? Faster setup, consistent high-quality sorting, and a significant increase in productivity.”
Conclusion
As the world adjusts to the impact of AI on our daily lives, pioneers are leveraging this monumental technological advancement to significantly enhance workflow, production, quality, and ultimately, profitability.
Defects manifest in countless ways across all cherries, and only AI possesses the real-time recognition capability to identify them with unprecedented precision. Having processed over 10,000 tons of cherries just this past season, packers throughout the Southern Hemisphere using gpVision with radiai report spectacular results. They have achieved more accurate and innovative defect identification and appreciate the system’s ease of use.
Article from: https://www.freshplaza.es/article/9714986/la-ia-es-el-futuro-de-la-clasificacion-de-cerezas/
Translation: SmartCherryworld
Image: SmartCherry.cl
