YOLO11 models tested for tree tomato maturity detection

Limits of manual inspection

Tree tomato (Solanum betaceum Cav.) is a fruit-bearing member of the Solanaceae family native to South America. Its fruit has nutritional, functional and economic value. Assessing maturity and deciding when to harvest, however, continue to depend primarily on manual visual inspection.

Under field conditions, this practice can be subjective and inconsistent. Different observers may judge the same fruit differently, making standardized harvest decisions more difficult. The study approached this problem through automation based on computer vision.

YOLO11 and smartphone images

The research aimed to evaluate the performance of YOLO11-based object detection models for automatically identifying the maturity of Solanum betaceum fruit. The assessment was designed around a real-world agricultural setting rather than being limited to controlled laboratory conditions.

The dataset used for the models contained 200 images captured with smartphones. The available excerpt and summaries do not specify the phone models, image locations, maturity categories or annotation procedure. They also do not explain how the images were divided into training, validation and test subsets, or identify the exact YOLO11 variants included in the evaluation.

These missing methodological details limit what can be concluded from the supplied material about the scope and reproducibility of the work. The available information establishes the research objective and dataset size, but it does not provide a complete account of model development or testing.

Performance details not provided

According to the supplied material, the study evaluated model performance, but no numerical results are included. It does not report accuracy measures, error rates, processing speeds or comparisons with alternative methods. The provided information therefore does not show how reliably the models distinguished among maturity states.

The study illustrates how smartphone imaging and deep learning can be combined for an agricultural task intended to support harvest-related decisions. Its focus is tree tomato, and the supplied material does not discuss applications in grape growing or winemaking. Any assessment of its relevance to other fruit crops would require methodological and performance details beyond those provided here.

The original research was published by Frontiers in Plant Science.