Selective harvesting of greenhouse cucumbers requires accurate detection with low inference latency. In greenhouse canopies, mature cucumbers are often partly occluded and visually similar to surrounding stems and leaves, which makes harvestability recognition difficult. Existing real-time detectors still struggle to preserve fine boundary cues, capture long-range context, and remain compact enough for on-device inference under these conditions. This study proposes MCS-DETR, an efficient multi-scale context-aware detector built on RT-DETR. Instead of increasing model scale, MCS-DETR redesigns shallow feature extraction, high-level contextual interaction, and cross-scale feature aggregation within a compact framework. A shallow feature level is also retained to preserve fine contour information. On the greenhouse cucumber dataset, MCS-DETR achieved 93.4% mAP@0.5 and 76.8% mAP@0.5:0.95, outperforming RT-DETR-R18 while requiring fewer parameters and less computation. On an NVIDIA Jetson Orin NX Super (Hunan ChuangLebo Intelligent Technology Co., Ltd., Room 2003, Building C, Xinchanghai Digital Center, Changsha Economic and Technological Development Zone, Changsha, Hunan, China) platform, it reached 26.3 FPS after TensorRT acceleration. These results indicate that MCS-DETR can provide an efficient on-device perception module for real-time greenhouse cucumber detection.

Soil tillage is energy-intensive, and the cutting resistance of soil-engaging tools is its largest single loss. Ultrasonic vibration-assisted cutting (UVAC) can reduce this resistance, but quantitative studies have concentrated on longitudinal vibration; transverse vibration, which loads the soil by cyclic lateral squeeze and release, remains largely unexplored. A calibrated discrete element model of a transverse UVAC tool–soil system is developed to determine how amplitude, frequency and cutting speed govern drag reduction and to identify the contact-level origin of the effect. Bench tests at 20 kHz supplement the simulated design space with an additional operating condition and confirm the predicted forces at eight settings—five vibration conditions and three baselines—with a mean relative error of 8.7% and the baselines within 6%; one calibrated parameter set is used throughout. Over 30–70 µm, 25–45 kHz and 0.03–0.07 m s⁻¹, amplitude dominates, raising the drag reduction from 44.10% to 68.18%, whereas frequency acts mainly through its quadratic term and its interaction with amplitude, and cutting speed enters the same coupling. Box–Behnken optimization places the optimum at 70 µm, 25 kHz and 0.07 m s⁻¹, with a DEM-verified reduction of 69.25% against a predicted 69.53%. Contact analysis shows that the vibration shifts load from normal compression to tangential shear and converts bond failure from concentrated rupture into progressive degradation. Amplitude is therefore the primary design variable for transverse UVAC in sandy loam at 9% moisture, with frequency and speed modulating its expression.