cs.ROMar 24, 2026

Active Robotic Perception for Disease Detection and Mapping in Apple Trees

Authors: Hayden Feddock, Francisco Yandun, Srđan Aćimović, Abhisesh Silwal

Organizations: H. Feddock, F. Yandun, and A. Silwal are with the Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213, USA · S. Aćimović is with the Department of Plant Pathology, Virginia Polytechnic Institute and State University, Blacksburg, VA, USA

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Large-scale orchard production requires timely and precise disease monitoring, yet routine manual scouting is labor-intensive and financially impractical at the scale of modern operations. As a result, disease outbreaks are often detected late and tracked at coarse spatial resolutions, typically at the orchard-block level. We present an autonomous mobile active perception system for targeted disease detection and mapping in dormant apple trees, demonstrated on one of the most devastating diseases affecting apple today -- fire blight. The system integrates flash-illuminated stereo RGB sensing, real-time depth estimation, instance-level segmentation, and confidence-aware semantic 3D mapping to achieve precise localization of disease symptoms. Semantic predictions are fused into the volumetric occupancy map representation enabling the tracking of both occupancy and per-voxel semantic confidence, building actionable spatial maps for growers. To actively refine observations within complex canopies, we evaluate three viewpoint planning strategies within a unified perception-action loop: a deterministic geometric baseline, a volumetric next-best-view planner that maximizes unknown-space reduction, and a semantic next-best-view planner that prioritizes low-confidence symptomatic regions. Experiments on a fabricated lab tree and five simulated symptomatic trees demonstrate reliable symptom localization and mapping as a precursor to a field evaluation. In simulation, the semantic planner achieves the highest F1 score (0.6106) after 30 viewpoints, while the volumetric planner achieves the highest ROI coverage (85.82%). In the lab setting, the semantic planner attains the highest final F1 (0.9058), with both next-best-view planners substantially improving coverage over the baseline.

Figures & tables

Fig. 1 : Autonomous disease inspection system (left) and representative fire blight symptoms (right). Shoot blight instances exhibit a characteristic shepherd’s crook (top), while canker regions appear as sunken or darkened lesions on woody tissue (bottom).
Fig. 2 : Effect of our flash camera on a canker in the afternoon. Image on the left was captured using an iPhone 15 max and the image on the right was captured using the flash camera detailed in [ 4 ] which is the primary camera used in our robotic system.
Fig. 3 : The perception pipeline (top) generates semantic 3D reconstructions from stereo depth and instance segmentation. The viewpoint planning module (bottom) first generates candidate viewpoints using a plane-based sampling strategy (baseline planner), augments them with orientation variations, and then selects viewpoints by clustering either frontier voxels (volumetric planner) or semantically uncertain regions (semantic planner) and sampling poses oriented toward the cluster centroids.
Fig. 4 : Blender-rendered symptomatic apple trees used in simulated Gazebo evaluations. Each tree contains randomly distributed shepherd’s crooks and cankers to simulate realistic outbreak variability and geometric complexity.
Fig. 5 : Laboratory setup with the Husky mobile base, xArm6 manipulator, and flash stereo camera inspecting a fabricated symptomatic tree. The black backdrop shown is purely for cosmetic purposes and is not required for system operation.
ParameterValue
OctoMap resolution0.04 m
α\alpha (utility weight)0.1
β\beta (semantic weight)0.7
Camera max range0.9 m
Background confidence0.3
Detection confidence threshold0.3
TABLE I : Planning Parameters
 SimulationLaboratory
PlannerF1Coverage (%)F1Coverage (%)
Baseline0.512059.570.844870.93
Volumetric0.512585.820.867097.46
Semantic0.610674.340.905896.02
TABLE II : Planner performance at 30 viewpoints
Fig. 6 : Planner performance in simulation (top) averaged across five symptomatic trees and in the lab setting (bottom). Curves show mean F1 score and ROI coverage over the first 30 viewpoints; shaded regions denote standard deviation across runs.
More in this topic

Robotic Inspection

Robotic inspection uses autonomous or remotely operated robots to examine infrastructure, equipment, and environments, detect defects, and document condition.

See all 50 papers

Explore similar work

CardsList
Sep 1, 2026

SG-AMP: Scene-Graph-Guided Active Perception and Semantics-Aware Motion Planning for Pepper Plants

We present SG-AMP, integrating robust depth completion with input-conditioned uncertainty, persistent panoptic mapping, plant scene-graph reasoning, and semantics-aware active view-motion planning. Beyond inspecting uncertain observed regions, the scene graph explicitly hypothesizes unobserved pepper--peduncle attachments and directs close-range sensing toward them. Candidate views are selected according to expected information gain, while class-dependent motion costs distinguish protected peppers, peduncles, and stems from conditionally traversable foliage. On pepper data, the perception network achieves 55.27%55.27\% semantic mIoU, 38.67%38.67\% PQ, and 40.62 mm40.62\,\mathrm{mm} depth RMSE, while input-conditioned uncertainty improves NYUv2 NLL from −1.6518-1.6518 to −1.6925-1.6925 and AUSE from 0.01020.0102 to 0.00870.0087.

Rohit Menon, Shiva Rudra Lolla, Niklas Mueller-Goldingen +3

Humanoid Robots Lab, University of Bonn, Germany · Center for Robotics, Bonn, Germany · Institute of Geodesy and Geoinformation, University of Bonn +1

Depth CompletionAgricultural RoboticsMotion PlanningGraph ReasoningActive Visual PerceptionAdaptive Sensing
AskPDF
Apr 11, 2026

A Comparative Study of Modern Object Detectors for Robust Apple Detection in Orchard Imagery

Accurate apple detection in orchard images is important for yield prediction, fruit counting, robotic harvesting, and crop monitoring. However, changing illumination, leaf clutter, dense fruit clusters, and partial occlusion make detection difficult. To provide a fair and reproducible comparison, this study establishes a controlled benchmark for single-class apple detection on the public AppleBBCH81 dataset using one deterministic train, validation, and test split and a unified evaluation protocol across six representative detectors: YOLOv10n, YOLO11n, RT-DETR-L, Faster R-CNN (ResNet50-FPN), FCOS (ResNet50-FPN), and SSDLite320 (MobileNetV3-Large). Performance is evaluated primarily using COCO-style [email protected] and [email protected]:0.95, and threshold-dependent behavior is further analyzed using precision-recall curves and fixed-threshold precision, recall, and F1-score at IoU = 0.5. On the validation split, YOLO11n achieves the best strict localization performance with [email protected]:0.95 = 0.6065 and [email protected] = 0.9620, followed closely by RT-DETR-L and YOLOv10n. At a fixed operating point with confidence >= 0.05, YOLOv10n attains the highest F1-score, whereas RT-DETR-L achieves very high recall but low precision because of many false positives at low confidence. These findings show that detector selection for orchard deployment should be guided not only by localization-aware accuracy but also by threshold robustness and the requirements of the downstream task.

Mohammed Asad, Ajai Kumar Gautam, Priyanshu Dhiman +1

Dept. of Electronics and Communications Engineering, Delhi Technological University, New Delhi, India

Object DetectionAgricultural Image Analysis
AskPDF
Sep 22, 2026

ArborSplat: Online Semantic Gaussian Splatting SLAM for Orchards

Orchard robots need maps that preserve small but semantically important structures such as trunks, trellises, and fruit. 3D Gaussian Splatting (3DGS) SLAM achieves high photometric fidelity. However, its optimization remains appearance-driven, and transferring image semantics to 3D points is unreliable for thin structures, whose pixels may receive depth from background surfaces. We present ArborSplat, an online semantic 3DGS SLAM system that tracks with LiDAR odometry and optimizes semantics directly on the Gaussian map, constrained by class-specific height bands above a ground plane fitted to each keyframe's stereo point cloud, and fuses multi-view evidence into a semantic point cloud online while rejecting labels inconsistent with the local ground surface or with monocular depth. Class-constrained refinement reserves Gaussian capacity for underrepresented structures and, under reduced budgets, increases training-view accuracy on tree classes. We evaluate the approach on apple and pear orchards during dormancy, flowering, and harvesting. On full routes, it keeps ATE below 0.5 m on all 12 traversals. On shared 301-frame segments, it exceeds SGS-SLAM and GS3LAM by 0.23 to 0.50 training-view and 0.15 to 0.36 held-out mIoU while running 1.7 to 7.5 times faster, whereas SemGauss-SLAM runs out of GPU memory on all six.

Alessandro Masini, Matteo Frosi, Mirko Usuelli +1

Dipartimento di Elettronica, Informazione e Bioingegneria (DEIB), Politecnico di Milano, 20133 Milano, Italy

3DGS SLAM3D Semantic SegmentationAgricultural Robotics
AskPDF