Trunk diameter is a valuable parameter for characterizing tree size, vigor, and growth, but its determination at commercial scale still requires considerable time and labor.
A recent study conducted in the United States on tart cherry developed and validated a low-cost mobile system capable of automatically estimating trunk diameter and associating it with the position of individual trees within the orchard.
The approach integrates RGB-D cameras, deep-learning algorithms for trunk detection and segmentation, three-dimensional reconstruction, and a driveline-based odometry system, avoiding reliance on GNSS, whose accuracy can decrease substantially beneath the tree canopy.
The system was installed on a utility vehicle equipped with two lateral RGB-D cameras capable of acquiring images during normal travel along the orchard rows.
The images were processed through a pipeline using YOLOv11 to identify the trunk closest to the camera and the Segment Anything Model to segment its surface.
Depth-associated pixels were converted into three-dimensional coordinates, allowing the trunk geometry to be reconstructed and its diameter to be determined at 40 cm above the ground, corresponding to the height used for manual measurements.
Validation was initially conducted in an experimental orchard of approximately 1 ha containing 462 mature ‘Montmorency’ trees grafted onto Mahaleb rootstock, and subsequently in a 9.5 ha commercial orchard containing approximately 3,600 trees.
Under unobstructed viewing conditions, the system showed good agreement with manual measurements, with a mean absolute error (MAE) of 0.95 cm, an RMSE of 1.16 cm, and a coefficient of determination (R2) of 0.79.
When all 462 trees in the experimental block were automatically evaluated by averaging multiple observations per tree, the MAE was 1.40 cm. The error decreased to 1.10 cm after excluding trees characterized by particularly severe occlusion.
The result is particularly interesting because it was obtained under real operating conditions characterized by variable illumination, basal vegetation, branches, leaning trunks, and, especially, bark deformation and scars associated with mechanical harvesting.
The redundancy of image acquisition proved to be essential: although only a proportion of frames showed the trunk under optimal conditions, each tree generally had multiple useful observations, allowing the estimate to be stabilized through averaging.
For along-row geolocation, digital odometry based on signals from the vehicle drivetrain was developed instead, replacing GNSS because of the positional drift observed under the canopy.
The system was subsequently applied across the entire commercial orchard, where trunk-diameter maps revealed spatial patterns consistent with those of canopy structure obtained from UAV surveys. Areas characterized by trees with larger trunks tended to correspond to areas with greater canopy height.
Although it does not necessarily replace higher-precision technologies such as LiDAR, the system offers an important advantage in terms of cost and simplicity: the entire platform can be assembled from commercially available components for approximately US$1,000–1,500 and installed on a vehicle already used for routine orchard operations.
The main areas for further improvement concern the detection of heavily occluded trunks, system autonomy, and positioning accuracy, which could be further enhanced through RTK-GNSS or inertial sensors.
Overall, the study demonstrates that trunk diameter can become an automatically acquired parameter and be transformed into a spatial map of the entire orchard, providing a basis for defining homogeneous management zones and developing decision-support systems for precision tart cherry production.
Source: Wedegaertner, K., Yost, K., Safre, A., Black, B., Young, S., & Torres-Rua, A. (2026). Integrating mobile RGB-D imaging and digital odometry for trunk diameter mapping in tart cherry orchards. Smart Agricultural Technology, 102140. https://doi.org/10.1016/j.atech.2026.102140
Image source: Stefano Lugli
Andrea Giovannini
PhD in Agricultural, Environmental and Food Science and Technology - Arboriculture and Fruitculture, University of Bologna, IT
18 Jun 2026
France’s 2026 cherry production has been revised down after hail in Auvergne-Rhône-Alpes, weak demand and early Spanish competition. Prices remain under pressure while Europe expects a solid harvest, with Italy recovering and trends in Occitania and Provence-Alpes-Côte d’Azur.
17 Jun 2025
Calatayud and Aranda cherries gain technical support from the Zaragoza Provincial Council for PGI recognition. A key step in a journey started in 2018, backed by local institutions, growers, and the Aragonese Government to promote this outstanding fruit.
11 Sep 2026
Artificial intelligence is transforming fruit growing in Chile: computer vision, machine learning and data help measure quality, yields and waste, reduce errors and improve decisions from the orchard to packing lines and export markets, including cherries and other fruit crops.
11 Sep 2026
In Apulia, HTS identified Pomes virus Greece (PVGR) in sweet cherry for the first time in Italy. The virus was detected in 8.7% of the samples analysed, while one case of Cherry green ring mottle virus also marks the first regional report of CGRMV in southern Italy.
