To address the lack of bird perception, ineffective output, and sound-source habituation in conventional orchard acoustic bird-repellent devices, this study develops an image-recognition-triggered acoustic bird-repellent prototype system based on an improved MobileViT-S. The system discriminates between “bird” and “no bird” in orchard images and controls an STM32 terminal via TCP Socket to output audible sound and variable-frequency acoustic waves on demand. To improve performance under small-target, foliage-occlusion, and complex-background conditions, multi-scale feature fusion and the CBAM attention mechanism are introduced into MobileViT-S. Ablation experiments show that the improved model achieves an Accuracy of 93.9% and an F1-score of 92.6%. Eighteen days of field operation show that the system can continuously complete image acquisition, bird-presence discrimination, command transmission, acoustic-terminal start/stop control, and switching among three operating modes, thereby verifying the operational feasibility of the system. Because no concurrent control, randomization, or independent replication was included, and because the three acoustic modes were implemented sequentially over time, the present results can only demonstrate the operational feasibility of the prototype system and should not be interpreted as evidence of bird-repelling efficacy or fruit-damage reduction. These effects require further validation through controlled field trials.

Accurate fruit instance segmentation is the premise of intelligent picking, precision spray and phenotypic parameter measurement. A green pepper fruit instance segmentation method based on SDBnet in near color environment is proposed in this paper. Based on the mutual constraints between the semantic representation ability and geometric information representation ability of feature maps at different levels, a network architecture design method suitable for effective feature map extraction based on the characteristics of target size is proposed. Specifically for elongated green pepper targets, a Shallow Feature Enhancement Network Architecture (SFENA) is proposed based on this method. The proposed Dual View Channel Attention Block (DVCAB) utilizes dual view channel attention created by pooling mean view and convolutional weight view, as well as convolution branch and optional residual branch construction, to achieve the extraction of multi-dimensional fruit effective feature information. A Binary Distribution Focal Loss (BDFL) is proposed based on the idea of probability distribution, which optimizes the probability of the left and right positions near the binary segmentation label, so that the segmentation positions focus on the vicinity of the label value. Multiple Population Genetic Algorithms (MPGA) optimizes the weight coefficients α and β of BDFL to achieve the optimal comprehensive accuracy of the model. The results of the field near color environment test showed that the Mask mAP50 and mAP50:95 for SDBnet fruit instance segmentation were 81.6% and 49.1%, which were 4.2% and 4.5% higher than the baseline model, respectively. The SDBnet parameter size is only 7.9 M, and the detection speed is 56.8f/s. Compared with the other 10 mainstream instance segmentation models, SDBnet has the highest accuracy performance indicators (Box/Mask mAP50 and mAP50:95), and the model has the smallest parameters and weight sizes. Therefore, SDBnet is the current State-of-the-Art model for green pepper fruit instance segmentation, which willprovide technical support for intelligent planting of green peppers.

Accurate and rapid fruit detection was very important for robot picking precisely, so the large model size and slow detection speed of the detection algorithm are problems that need to be solved urgently. An improved lightweight Faster R-CNN based on MobileNetV3 was proposed in this paper, which was used to detect fruits on Ori_RGB and Rb_RGB image datasets that collected by RGB-D camera in densely planted commercial pitaya orchards. On the Rb_RGB image datasets, the detection AP of 0.929 and 0.898 were obtained using MobileNetv3_large_FRCNN and MobileNetv3_small_FRCNN, which decreased 1.38% and 4.67% than that using VGG16_FRCNN respectively, and the detection time was 35.4 and 18.8 ms per image, which decreased 46.5% and 71.6% than that using VGG16_FRCNN respectively. On the Ori_RGB image datasets, the detection AP of 0.911 and 0.856 were obtained using MobileNetv3_large_FRCNN and MobileNetv3_small_FRCNN, which decreased 2.15% and 8.06% than that using VGG16_FRCNN respectively, and the detection time was 35.2 and 19.5 ms per image, which decreased 47.2% and 70.8% than that using VGG16_FRCNN respectively. Weight sizes of MobileNetv3_large_FRCNN and MobileNetv3_small_FRCNN were 3.19%, 1.15% of that of VGG16_FRCNN respectively. The detection AP values on the Rb_RGB image test set using three networks than that on Ori_RGB image test set increased 1.98%, 4.91%, and 1.18%, but image type had no significant effect on AP. The improved lightweight Faster R-CNN based on MobilenetV3 is expected to deploy to the embedded system of the fruit picking robot to detect pitaya, which would promote the development of robot picking technology.

Every year, up to 40% of the crops in the world are lost to pests. Plants have suffered from prolonged biotic stresses and abiotic stresses, which cause significant changes in complex crop ecosystems, necessitating intensive pest management strategies that have often been accompanied by the struggle against plant pests. Plant pests and diseases control methods heavily reliant on chemical pesticides have caused many adverse effects. One innovative method involves using ultrafine bubble (UFB) waters, which can enable pesticide reduction action for the plant pest control. The classification and six properties of UFBs were summarized, and the generation approaches of UFBs were introduced based on physical and chemical methods. The applications of UFBs and ozone UFB waters in plant protection practices were comprehensively reviewed, in which UFB waters against the plant pests and the soilborne, airborne and waterborne diseases were analyzed, and the abiotic stresses of crops in high-salinity soil and contaminated soil, drought, and soil with heavy metals were reviewed. Despite promising applications, UFB technology has limitations. Aiming at pesticide reduction and replacement using UFB waters, the mechanism of UFB water controlling plant pests and diseases, the molecular mechanism of UFB water affecting plant pest resistance, the plant growth in harsh polluted environments, the UFB behavior with hydrophobic and hydrophilic surfaces of crops, and the building of an integrated intelligent crop growth system were proposed.

To address the challenge of quickly tuning the resonance frequency of ultrasonic transducers in real-time after manufacturing, this paper proposes integrating Terfenol-D into the horn of a piezoelectric ultrasonic transducer. In practical applications, the transducer is mechanically excited by a piezoelectric vibrator connected to an alternating current power supply, while the Terfenol-D is surrounded by a coil connected to a direct current power supply. By utilizing the delta-E (ΔE) effect of Terfenol-D, the modulus of the material changes in response to variations in the surrounding magnetic field, enabling real-time online adjustment of the resonance frequency. The resonance frequency and amplification coefficient of the transducer were analyzed using the transfer matrix method. Then, an impedance analyzer was employed to measure changes in the transducer's impedance and resonance frequency under different electric currents. Experimental results demonstrate that the resonance frequency of the transducer increases with increasing electric current. The developed piezoelectric ultrasonic transducer, featuring Terfenol-D embedded in the horn, achieves rapid and real-time resonance frequency adjustment within a specific range. This innovation provides a novel solution for complex ultrasonic application scenarios requiring frequent transducer replacements.

Plants are subjected to long-term biotic stresses and abiotic stresses which cause sig-nificant changes in complex crop ecosystems and have been often accompanied by the struggle against plant pests. Pests control methods relying heavily on chemical pesti-cides have resulted in numerous adverse effects. One of innovative methods involves using ultrafine bubble (UFB) waters may realize the pesticide reduction action for the plant pest control. The classification and six properties of UFBs were summarized, and the generation approaches of UFBs were introduced based on physical and chemical methods. The applications of UFBs and ozone UFB waters in plant protection practices were comprehensively reviewed, in which, UFB waters against the plant pest insects and the soilborne, airborne and waterborne diseases were analyzed, and the abiotic stresses of crops in salinity soil and contaminated soil were reviewed. Because UFB water is not omnipotent, several prospects were proposed aiming at pesticide reduc-tion and replacement, for example, the mechanism of UFB water controlling plant pests and diseases, the molecular mechanism of UFB water affecting plant pest resistance, the plant growth in harsh polluted environments, the UFB behavior with hydrophobic and hydrophilic surfaces of crops and the building of integrated intelligent crop growth system.

This research introduced an innovative composite piezoelectric ultrasonic transducer with magnetically-tuned, enabling dynamic frequency modulation. The proposed configuration integrates three key components: a longitudinal sandwich-type piezoelectric transducer, a magnetically-tuned resonance horn assembly, and a conical horn. An electromechanical equivalent circuit model was established to analyze the transducer's vibration characteristics. Theoretical analysis, finite element simulation using COMSOL, and experimental validation were conducted to evaluate the resonance frequency and vibrational modes under different conditions. The transducer incorporates Terfenol-D, a smart magnetostrictive material exhibiting variable Young's modulus characteristics under externally applied DC magnetic fields through the Delta E-effect. This magneto-elastic coupling mechanism facilitates precise electromechanical resonance adjustment, achieving real-time frequency modulation within the composite transducer system. The experimental results show that as the current increases from 0.2 A to 5.2 A, the magnetic field increases from 18.58 Gs to 310.84 Gs, and the frequency of the transducer increases by 380 Hz. Experimental results demonstrate that the transducer achieved adjustable resonance frequencies in real-time. This means that the research has potential value for applications such as pesticide spraying, fertilizing, irrigation, fuel atomization, wet desulfurization, dust suppression and atomized drug delivery. This research provides a foundation for the development of advanced ultrasonic transducers with tunable frequencies for diverse industrial and environmental applications.

In modern agriculture, plant protection is the key to ensuring crop health and improving yields. Intelligent pesticide prescription spraying (IPPS) technologies monitor, diagnose, and make scientific decisions about pests, diseases, and weeds; formulate personalized and precision control plans; and prevent and control pests through the use of intelligent equipment. This study discusses key IPSS technologies from four perspectives: target information acquisition, information processing, pesticide prescription spraying, and implementation and control. In the target information acquisition section, target identification technologies based on images, remote sensing, acoustic waves, and electronic nose are introduced. In the information processing section, information processing methods such as information pre-processing, feature extraction, pest and disease identification, bioinformatics analysis, and time series data are addressed. In the pesticide prescription spraying section, the impact of pesticide selection, dose calculation, spraying time, and method on the resulting effect and the formulation of prescription pesticide spraying in a certain area are explored. In the implement and control section, vehicle automatic control technology, precision spraying technology, and droplet characteristic control technology and their applications are studied. In addition, this study discusses the future development prospectives of IPPS technologies, including multifunctional target information acquisition systems, decision-support systems based on generative AI, and the development of precision intelligent sprayers. The advancement of these technologies will enhance agricultural productivity in a more efficient, environmentally sustainable manner.

A machine vision and neural network-based method for the quantitative detection of hopper discharge characteristics based on the discharge time distribution is proposed. Glass beads and quartz sand were utilized as test objects. The prediction model of the relationship between the particle mass and the pixel value of the image was established by an artificial neural network, and the mass flow rate (MFR) was calculated via image prediction. The average relative errors of the predicted MFR for glass beads and quartz sand were found to be -1.31 % and -2.02 %, respectively. Based on the particle marking method, a convolutional neural network was used to classify the image according to whether there were marked particles in the image, and the mass flow index (MFI) was calculated after error correction. The average relative errors of the predicted MFI values for glass beads and quartz sand were found to be -1.43 % and 0.82 %, respectively.

In a densely planted orchard, factors such as light variation, branch occlusion, and fruit in non-picking rows had a great impact on the pitaya detection accuracy. In this study, a new WGB-YOLO network was developed and tested for multi-class pitaya fruits detection in target picking rows. The proposed WFE-C4 module was obtained by adding two wings feature enhancement structure based on Bottleneck and cascading MetaAconC functions, which independently enhanced feature extraction from the channel and spatial dimensions. A backbone network with WFE-C4 to replace YOLOv3 ' s Darknet53 was constructed. The proposed GF-SPP used average pooling and global average pooling instead of 2 maximum pooling in SPP, and the global average pooling features were used as independent channels to strengthen the average and maximum pooling features respectively, which simultaneously achieved multi-scale fusion of features and feature enhancement. The new WGB-YOLO network used a Bi-FPN structured head network to achieve a balanced fusion of multi-scale features. The tests showed that the mAP of multi-lass pitaya in the target picking rows was 86.0% using WGB-YOLO detection, while the AP of NO, FCC, and OB fruit were 96.0%, 84.4%, and 77.6%, respectively. WGB-YOLO improved the AP of the original model for detecting OB fruits by 10.5%, which indicated a significant improvement in model detection performance. Compared with 8 other deep networks such as YOLOv7, WGB-YOLO obtained the highest mAP for detecting multi-class pitaya while maintaining a better detection speed. WGB-YOLO showed good performance in detecting pitaya in densely pitaya planted orchards, which provided a technical foundation for fruit detection in robotic picking of the target rows.

BACKGROUND:Accurate pesticide inline mixing uniformity (PIMU) evaluation for direct nozzle injection systems (DNIS) helps evaluate system performance and develop efficient inline mixers. Based on supervised machine learning (ML), inline mixing images and computational fluid dynamics (CFD) simulations are directly associated for realizing intelligent PIMU predictions.RESULTS:Image sets can be reduced to less than 3% of the data size at the same time as retaining 98% of information using principal component analysis (PCA). The CFD results, as referenced values for ML, were justified by mixture sampling experiments. Enhanced images for the long-mixing tube effectively trained models including generalized linear model (GLM), support vector regression (SVR), BP-neural network (NNW), and classification and regression trees (CART). By testing the re-collected images, the verification accuracy of GLM was less than 95% and it failed to recognize uniformity differences under varying working conditions, whereas NNW, CART and SVR realized it with an accuracy for NNW and CART higher than 97% and for SVR slightly lower than 97%. By testing images of the jet mixer, the prediction accuracy compared with the CFD results of NNW and CART was also higher than 97%, although that for SVR was relatively lower, and insignificant declines in accuracy were observed on comparing the results with mixture sampling experiments.CONCLUSION:PCA facilitates evaluations of CFD-referenced PIMU using image-based ML. Models trained by enhanced image sets of the long-mixing tube have satisfactory performance. NNW and CART performed slightly better than SVR, and they can be used as tools to improve the rationality when evaluating PIMU in DNIS. © 2022 Society of Chemical Industry.

Rapid and accurate detection of green peppers are essential for their growth monitoring, yield estimation, phenotypic monitoring, and robotic harvesting. In order to simplify the detection model and improve the detection efficiency, the NSGA-II-based (Non-dominated Sorting Genetic Algorithm-II-based) pruning algorithm was proposed to obtain an optimal pruning that balanced the detection accuracy and speed of the pruned model. A green pepper detection model was trained using YOLOv5l, and the NSGA-II-based pruning algorithm was implemented to obtain a YOLOv5l pepper detection model. The number of model parameters, model size, and GFlops of the pruned model were reduced by 73.9 %, 73.5 %, and 62.7 %, respectively. The mAP0.5 of the pruned model was 81.4 %, only slightly lower (by 0.973 %) than that of the original model.The detection speed of the pruned model was 70.9f/s, which was 59.0 % higher than that of the model before pruning. The NSGA-II-based pruning also significantly outperformed other two algorithms, namely, Slim pruning and EagleEye pruning, in terms of number of parameters, model size, GFlops, and detection speed, with a slight reduction in mAP0.5 0.973 % compared to EagleEye pruning. Finally, the NSGA-II-based pruned YOLOv5l pepper detection model was compared with other 11 deep learning models. Except that the mAP0.5 was only 0.367 % lower than that of YOLOv4, our method again showed had obvious advantages in terms of parameter quantity, model size, GFlops, mAP0.5, and detection speed. This research provided a new method and insights for the pruning of deep learning models, which is a necessary step to deploy them in compact mobile devices for real-time applications.

An experimental method based on the discharge time distribution (DTD) and the finite element method (FEM) simulation was applied to quantitatively evaluate the flow pattern, mass flow rate (MFR), and coefficient of variation (CV) of the MFR of a granular flow during the discharge of conical and hyperbolic hoppers with different initial inclination angles. The mass flow index (MFI) and MFR of the conical and hyperbolic hoppers appear to increase with the initial inclination angle. However, the MFI of the conical hopper exceeds that of the hyperbolic hopper while the MFR of the conical hopper is lower than that of the hyperbolic hopper. The CV of the MFR decreases as initial inclination angle increases for conical hoppers, but is less affected by the initial inclination angle for hyperbolic hoppers.

In this review, through reviewing the history of the struggle between human beings and plant diseases, insects and weeds, more specifically thoughts on plant protection in ancient Chinese agricultural books, the recognition of plant pests as a target and six types of plant protection methods and 36 subdivision measures are summarized. Then, we focus on the development overview of pesticide application technology and conduct a systematic review by combining the development timeline of pesticide application and key technologies including performance measurement and the simulation and modeling of pesticide-spraying systems. Finally, three suggestions for further research are proposed from the perspectives of human beings’ and environmental health, sustainable and eco-friendly application media and efficient application equipment systems in plant protection.

作为陆地森林生态系统骨架的树木长期遭受有害生物胁迫,危害林木正常生长,甚至造成林木死亡,对林业可持续发展带来严重威胁,为了实现对林业有害生物靶标及灾害区域的快速、准确检测和监测,需要识别获取林业有害生物防治靶标对象的特征.本研究首先重点围绕生物胁迫之植物病害、植物虫害、鼠(兔)危害、有害植物等直接靶标及其危害的间接靶标的主要特征分析,综述国内外采用不同途径检测和识别林业有害生物防治靶标的技术和装备研究概况.然后建议组织跨学科力量,探索研究基于植物胁迫预警靶标识别、传感网络协同防御控制的林业有害生物防治靶标识别技术及综合防治智能化系统.展望林业有害生物防治靶标识别被动传感技术和主动传感技术的未来发展,提出根据植物在生物胁迫下的物理化学特征表现开展机器视觉、气敏、力敏、热敏、声敏等防治靶标的被动传感技术研究,利用激光雷达扫描、光声效应主动引诱以及电子昆虫和气味主动引诱等主动传感技术,以解决被动传感难以识别特定有害生物靶标的问题,以及研发设计生物、磁敏、湿敏、热敏等新型传感器和研究超声波、LiDAR、辐射、卫星图像等多传感器融合技术,提高林业有害生物靶标的识别精度.最后围绕植物根系、野生哺乳动物、土栖白蚁、有害植物等特定靶标提出相应的靶标感知识别技术研究建议.

Spraying chemical pesticides is one of the important means to control plant pest, and the profile variable spraying is an important technology to achieve precise pesticide application. A profiling tracking control method and an improved algorithm based on CMAC-PID (Cerebellar Model Articulation Controller- Potential Induced Degradation) were proposed in the paper. The test results of the sprayer profiling tracking of the tree canopies showed that the profiling control system using the improved algorithm had significantly better dynamic tracking performance, and the overall mean tracking error was reduced by 35.0%, compared with the traditional CMAC-PID. A spray flow calculation method based on tree canopy volume and leaf area density was proposed. Outdoor testing of the profile variable spraying and conventional spraying was carried out. There was no significant difference between the two spraying methods in terms of droplet coverage, VMD (Volume Median Diameter), NMD (Number Median Diameter), spray quality parameter and relative span coefficient, as well as droplet deposition density. The spray coefficient of variation was reduced by 25.9% and 21.9% inside and outside the tree canopy, respectively. The mean value of the ground deposition coverage of the profile variable spraying and the traditional spray was 13.0% and 33.2%, respectively, indicating a significant impact on the ground droplet deposition coverage by the two spraying methods. The spray flow rate of the profile variable spraying could be decreased by 32.1% compared to the conventional spraying. Profile variable spraying would reduce the cost associated with pesticide use and environmental pollution.

针对混合试验图像所得均匀性指数计算结果难以直接匹配于被广泛认可的数值仿真参考值的问题,本文基于线性模型方法,将混合试验图像处理与数值仿真结果进行映射,在黏性水溶性农药与水在长直混合管内进行在线混合的试验条件下构建对应的线性预测模型,并采用射流混药器在线混合图像及仿真结果对上述模型进行检验.研究结果表明:不同图像方法(灰度直方图二阶矩(HSM)、改进面积加权法(OAU)、主成分分析法(PCA))对应最优线性拟合阶数不同,采用单独图像方法构建模型时最优阶次为4,决定系数R2高于0.95,采用2种图像方法组合和3种图像方法组合时最优阶次可分别降至3阶和2阶,R2则接近或高于0.98;载流流量Q为800～2 000 mL/min、混合比P为0.01～0.10条件下,基于HSM、OAU、PCA和线性模型,可实现实际混药器均匀性预测,所有模型预测误差均小于0.05,且采用一元和二元线性模型使得平均预测误差分别降低84.1％和79.8％,不同算法间预测结果极差分别降低31.6％和78.0％;采用基于PCA或OAU算法的一元模型进行预测时误差可控制在0.03以内,其精度高于不同算法组合预测的结果;采用基于HSM-PCA等算法组合的二元模型误差虽稍高于0.03,但也可避免单一图像指标计算不准确带来的预测风险.通过构建图像处理-数值仿真之间的映射关系,可为基于图像处理进行农药在线混合均匀性评估提供更加可行和合理的方法.

The performance of the profiling tracking control system was one of the important factors affecting the spray deposition and drift characteristics of the profiling variable spray. This paper proposed a dynamic threshold function to improve the CMAC-PID control algorithm to improve the real-time response speed and robustness of profiling tracking control system. The optimal parameters of dynamic threshold function was determined through a response surface experiment and optimal calculation. The continuous profiling tracking experiment results for outdoor tree canopy within 0-2.5 s showed that, compared with the CMAC-PID control algorithm, when the improved CMAC-PID control algorithm adopted by the profiling control system, the rise time was respectively shortened by 71.5%, 66.1% and 67.3%, the adjustment time was respectively shortened by 67.6%, 57.6% and 50.0%, and the overshoot was respectively 0.658%, 0.552% and 3.46% for the profiling angle position step response of the profiling mechanism module A_up, B and A_down. The continuous profiling tracking experiment results for outdoor tree canopy within 2.5-11.5 s showed that, compared with the CMAC-PID control algorithm, when the improved CMAC-PID control algorithm was adopted by the profiling control system, the profiling tracking response curves and the profiling target angle curves were closer to coincidence, and the average error of profiling tracking has been reduced by 35.9%, 57.4% and 38.1% respectively for the profiling mechanism module A_up, B and A_down. It showed that the profiling control system using the improved CMACPID algorithm had better response speed and profiling tracking performance.

农林植物受生物及非生物胁迫的发生面积居高不下,化学农药防治是消灭或控制病虫草害等有害生物胁迫的最主要方法,而农药喷雾性能直接影响着病虫草害防治效果.通过农药喷雾全过程性能综合分析图,归纳了农药喷雾过程中的雾化性能、输运沉降性能和沉积性能.其中,雾化性能包括雾流锥角形状、射流贯穿长度、液膜破碎距离以及药液在线混合等喷雾宏观特性和雾滴变形、分裂、聚并、碰撞和雾滴尺寸、雾滴速度等喷雾微观特性;雾滴形成后到达靶标前的流场(电场)性能及其雾滴在流场(电场)中的行为是输运沉降性能的重要测试内容;雾滴到达靶标后包括浸润、持留、蒸发、弹跳、滑落等雾滴沉积性能.综述了农药喷雾性能综合测试方法、关键喷雾部件性能测试、雾化过程性能测试、输运过程喷雾流场及沉降性能测试、沉积过程雾滴运动行为测试、防治效果测试、农药残留测试等研究概况,提出了开展系列化喷雾性能标准试验系统研究、模块化田间喷雾性能测试仪研制、综合喷雾性能测试系统、农药喷雾模型与喷雾性能物理测试互动耦合、生物农药喷雾性能测试、农药残留快速智能化检测、根据靶标植物的表型特征调节农药用量的测试技术研究及智能化无人农场植保作业性能测试系统设计等建议,以形成系统化农药喷雾性能测试体系.