Weipeng Zhang

Conventional orchard spraying often suffers from a mismatch between spraying parameters and the spatial structure of fruit tree canopies, resulting in uneven spray deposition and low spray utilization efficiency. To address these issues, this study developed a canopy feature-based profile variable-rate orchard sprayer. LiDAR was used to acquire spatial information on fruit tree canopies, based on which a mapping relationship between canopy profile characteristics and profile adjustment parameters was established. A profile adjustment mechanism consisting of a lifting platform, a translation mechanism, and a five-section profile spraying arm was designed to enable adaptive adjustment of nozzle positions and spray directions according to canopy morphology. In addition, a variable-rate spraying model was developed based on the effective canopy volume to achieve dynamic matching between spray volume and canopy spatial distribution. Field experiments were conducted in a high-spindle apple orchard to evaluate the performance of the developed sprayer. The results showed that, under the canopy-adaptive spraying mode, the average droplet coverage rate on the adaxial leaf surface was 37.65%, with a vertical coefficient of variation of 2.83%. The horizontal coefficients of variation for the upper, middle, and lower canopy layers were 11.50%, 8.03%, and 5.73%, respectively. Compared with the fixed-position quantitative spraying mode, the canopy-adaptive spraying mode reduced ground spray loss by 69.47% and water consumption by 51.93%. Overall, the canopy-adaptive spraying mode improved droplet deposition uniformity, reduced ground spray loss, and decreased spray volume, demonstrating its potential for precision plant protection in orchards.

To address the difficulty of real-time detection of seed-filling performance in pneumatic suction seed metering devices under high-speed operation—where seed targets are tiny, prone to adhesion, and affected by motion blur—this paper proposes a lightweight online detection algorithm, YOLOv8n-MA. First, according to the seed adsorption characteristics of the suction holes, the detection targets are divided into three categories: none, one, and two. Second, based on YOLOv8n, the backbone network is replaced with MobileNetV1 to reduce computational cost, and an ACmix attention module is integrated into the Neck to enhance feature representation for the three suction-hole states. Finally, to meet the demand for low-latency inference on resource-constrained devices, the model is deployed on an edge computing controller to achieve real-time detection. Experimental results show that, compared with the original YOLOv8n, the parameters and FLOPs of YOLOv8n-MA are reduced by 34.4% and 59.8%, respectively, while the mean average precision (mAP) is improved by 2.0% to 96.8%, achieving a superior trade-off between accuracy and efficiency over other detection models of the same category, such as YOLOv5n, YOLOv9n, and YOLOv10n. In field tests, the detection accuracy reaches 95.02% at 12 km/h and 92.65% at 15 km/h. The proposed method provides effective technical support for the intelligent monitoring and control of precision seeding under high-speed operation.

Full-condition monitoring and intelligent yield prediction and decision-making technology for wheat combine harvesters