Selected work

Representative Papers

Boundary-Enhanced Segmentation of Pig Point Clouds in Commercial Housing Environments

Aug 12, 2026

This study addresses the challenges of ambiguous boundaries, local adhesion, and erroneous segmentation between pigs and background in commercial pig barns by proposing a boundary-aware point cloud segmentation method. The approach employs an Octree Transformer backbone to effectively integrate fine-grained local geometric details with global semantic context. It introduces soft-distance boundary pseudo-labels for continuous boundary supervision and incorporates a novel bidirectional cross-boundary semantic module to explicitly model interactions between boundary cues and semantic features. Experimental results demonstrate that the proposed method significantly outperforms state-of-the-art models on a comprehensive dataset, achieving superior performance in segmentation accuracy, mean Intersection over Union (mIoU), and boundary delineation, thereby providing high-quality point cloud inputs for precision livestock farming.

0 citationsRead paper

Vision-Based Early Fault Diagnosis and Self-Recovery for Strawberry Harvesting Robots

Jan 05, 2026arXiv.org

This study addresses frequent harvesting failures in strawberry-picking robots caused by fragmented visual perception, misalignment between fruit and end-effector, empty grasps, and fruit slippage. To this end, the authors propose a fault diagnosis and self-recovery framework integrating multi-task visual perception with corrective control. They introduce SRR-Net, an end-to-end model that jointly performs strawberry detection, segmentation, and maturity estimation. Leveraging a micro-optical camera mounted on the gripper, the system incorporates a relative position error compensation mechanism between target and gripper, alongside a MobileNetV3-Small-based grasp state classifier and an LSTM-based slippage prediction module for early fault detection and real-time intervention. Experimental results demonstrate that SRR-Net achieves high performance in detection (precision: 0.895, recall: 0.813), segmentation (precision: 0.887, recall: 0.747), and maturity estimation (MAE: 0.035), with an inference speed of 163.35 FPS, significantly enhancing harvesting reliability and efficiency.

0 citationsRead paper

Recent publications

Latest Papers

Boundary-Enhanced Segmentation of Pig Point Clouds in Commercial Housing Environments

Aug 12, 2026

This study addresses the challenges of ambiguous boundaries, local adhesion, and erroneous segmentation between pigs and background in commercial pig barns by proposing a boundary-aware point cloud segmentation method. The approach employs an Octree Transformer backbone to effectively integrate fine-grained local geometric details with global semantic context. It introduces soft-distance boundary pseudo-labels for continuous boundary supervision and incorporates a novel bidirectional cross-boundary semantic module to explicitly model interactions between boundary cues and semantic features. Experimental results demonstrate that the proposed method significantly outperforms state-of-the-art models on a comprehensive dataset, achieving superior performance in segmentation accuracy, mean Intersection over Union (mIoU), and boundary delineation, thereby providing high-quality point cloud inputs for precision livestock farming.

0 citationsRead paper

Vision-Based Early Fault Diagnosis and Self-Recovery for Strawberry Harvesting Robots

Jan 05, 2026arXiv.org

This study addresses frequent harvesting failures in strawberry-picking robots caused by fragmented visual perception, misalignment between fruit and end-effector, empty grasps, and fruit slippage. To this end, the authors propose a fault diagnosis and self-recovery framework integrating multi-task visual perception with corrective control. They introduce SRR-Net, an end-to-end model that jointly performs strawberry detection, segmentation, and maturity estimation. Leveraging a micro-optical camera mounted on the gripper, the system incorporates a relative position error compensation mechanism between target and gripper, alongside a MobileNetV3-Small-based grasp state classifier and an LSTM-based slippage prediction module for early fault detection and real-time intervention. Experimental results demonstrate that SRR-Net achieves high performance in detection (precision: 0.895, recall: 0.813), segmentation (precision: 0.887, recall: 0.747), and maturity estimation (MAE: 0.035), with an inference speed of 163.35 FPS, significantly enhancing harvesting reliability and efficiency.

0 citationsRead paper