Abstract
Environmental perception in unstructured agricultural settings represents a fundamental challenge for autonomous orchard robotics. This study presents a real-time decision-level fusion framework integrating 32-channel LiDAR with a monocular camera for tree detection and localization in orchard environments. The perception system incorporates three key innovations: (1) triple-stage point cloud preprocessing utilizing region-of-interest segmentation, RANSAC-based ground removal, and intensity-based filtering to reduce computational load by 68%; (2) enhanced DBSCAN clustering with adaptive cluster merging and geometric shape filtering specifically optimized for tree morphology; and (3) decision-level fusion employing the Hungarian algorithm for optimal sensor association with semantic validation. A custom orchard dataset comprising 5000 annotated images was collected to train the YOLOv7 detection model, achieving 95.9% precision, 95.7% recall, and 75.2% mAP@0.5:0.95 on the test set. Experimental validation in a simulated orchard demonstrated superior performance: the fusion system achieved 97.80% precision and 96.50% recall, representing a 4.36% precision improvement over enhanced single-sensor approaches and a 29.44% gain vs baseline methods. Real-time processing capability was maintained at 41 ms per frame (24.4 FPS) with a 10-m effective perception range on moderate-cost hardware. Temporal robustness analysis across 100-frame sequences confirmed stable operation with 69% perfect detection, which is substantially higher than single-sensor methods. Field trials validated that the proposed LiDAR–camera fusion framework provides sufficient perception accuracy, robustness, and real-time performance to support practical autonomous navigation.
Abstract from the paper’s OpenAlex record; open copy licensed cc-by. Read at the source ↗
The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.
Claims 13
Claim · model-written summary
Point cloud preprocessing reduced computational load by 68%.
Verbatim
e0“triple-stage point cloud preprocessing utilizing region-of-interest segmentation, RANSAC-based ground removal, and intensity-based filtering to reduce computational load by 68%”
t0“orchard”
Claim · model-written summary
YOLOv7 achieved 95.9% precision on the orchard test set.
Verbatim
e0“achieving 95.9% precision, 95.7% recall, and 75.2% mAP@0.5:0.95”
t0“orchard”
Claim · model-written summary
YOLOv7 achieved 95.7% recall on the orchard test set.
Verbatim
e0“achieving 95.9% precision, 95.7% recall, and 75.2% mAP@0.5:0.95”
t0“orchard”
Claim · model-written summary
YOLOv7 achieved 75.2% mAP@0.5:0.95 on the orchard test set.
Verbatim
e0“achieving 95.9% precision, 95.7% recall, and 75.2% mAP@0.5:0.95”
t0“orchard”
Claim · model-written summary
The fusion system achieved 97.80% precision in a simulated orchard.
Verbatim
e0“the fusion system achieved 97.80% precision and 96.50% recall”
e1“Experimental validation in a simulated orchard”
t0“orchard”
Claim · model-written summary
The fusion system achieved 96.50% recall in a simulated orchard.
Verbatim
e0“the fusion system achieved 97.80% precision and 96.50% recall”
e1“Experimental validation in a simulated orchard”
t0“orchard”
Claim · model-written summary
Fusion precision improved by 4.36% over enhanced single-sensor approaches.
Verbatim
e0“a 4.36% precision improvement over enhanced single-sensor approaches”
t0“orchard”
Claim · model-written summary
Fusion precision gained 29.44% versus baseline methods.
Verbatim
e0“a 29.44% gain vs baseline methods”
t0“orchard”
Claim · model-written summary
The system processed each frame in 41 ms.
Verbatim
e0“Real-time processing capability was maintained at 41 ms per frame (24.4 FPS) with a 10-m effective perception range”
t0“orchard”
Claim · model-written summary
The system processed 24.4 FPS.
Verbatim
e0“41 ms per frame (24.4 FPS)”
t0“orchard”
Claim · model-written summary
The effective perception range was 10 m.
Verbatim
e0“10-m effective perception range”
t0“orchard”
Claim · model-written summary
Across 100-frame sequences, perfect detection reached 69%, above single-sensor methods.
Verbatim
e0“across 100-frame sequences confirmed stable operation with 69% perfect detection, which is substantially higher than single-sensor methods”
t0“orchard”
Claim · model-written summary
Field trials found fusion perception sufficient for practical autonomous navigation.
Not reportedComparator
Verbatim
e0“Field trials validated that the proposed LiDAR–camera fusion framework provides sufficient perception accuracy, robustness, and real-time performance to support practical autonomous navigation”
t0“orchard”