基于改进YOLOv5的茶叶杂质检测算法

针对现有目标检测算法检测茶叶杂质精度低、速度慢的问题,该研究提出了一种基于改进YOLOv5的茶叶杂质检测算法。采用K-Means聚类算法对杂质真实框聚类,以获取适合茶叶杂质特征的锚框尺寸;通过在主干特征提取网络CSPDarkNet中引入前馈卷积注意力机制（Convolutional Block Attention Module,CBAM）,将茶叶杂质输入特征图依次经过通道注意力模块和空间注意力模块,获得特征图通道维度和空间维度的关键特征;在颈部网络中添加空间金字塔池化（Spatial Pyramid Pooling,SPP）模块,融合并提取不同感受野的关键特征信息;将普通卷积替换成深度可分离卷积,增大小目标预测特征图的置信度损失权重,构建了轻量化的改进YOLOv5网络结构模型;分别制作了铁观音茶叶中混合有稻谷、瓜子壳、竹片和茶梗4种杂质的数据集并进行茶叶杂质检测试验。结果表明,改进的YOLOv5比常规YOLOv5在茶叶杂质检测中具有更高的置信度分数,且定位更为准确,未出现漏检现象。改进YOLOv5的多类别平均精度（Mean Average Precision,mAP）和每秒传输帧数（Frame Per Second,FPS）达到96.05%和62帧/s,均优于主流的目标检测算法,验证了改进算法的高效性和鲁棒性。该研究成果可为提升茶叶制作过程中小目标杂质检测精度与检测速度奠定基础。     

深潭的功能及在城市河道治理中的构建模式

为充分了解自然界河流深潭的成因和功能,选取山东诸河部分河段,对河流深潭的水量状况、底质、水生动物调查进行,结合相关文献总结深潭形成的主要原因,阐述深潭在生态、经济、景观方面的功能.考虑到城市河流的现状,认为人为地在城市河道修建深潭的方式,对城市河流的生态系统、河流水质和景观有着积极的作用.在相关研究的基础上提出深潭构建的5条原则,以及“浅滩-深潭”和“跌水-深潭”2种构建形式,旨在为城市河道的生态改造提供参考和借鉴.     

Effects of deep placement of fertilizer on periphytic biofilm development and nitrogen cycling in paddy systems

Periphytic biofilms are commonly presented at the water-soil interface in paddy fields. Different fertilization methods can affect the concentration and distribution of nutrients in paddy fields and thus affect the development of periphytic biofilms. In this study, the roles of periphytic biofilms in nitrogen(N) cycling in paddy systems and how they are affected by different fertilization methods were studied using microcosm experiments. Microcosms were prepared using soil samples from a paddy field and treated with surface and deep fertilization under light and dark conditions. Surface fertilization under light condition promoted the development of periphytic biofilms, while deep fertilization under dark condition inhibited their development. The development of periphytic biofilms increased the pH and dissolved oxygen levels in the overlying water. Surface fertilization resulted in high N concentrations in the overlying water and the topsoil layers, which enhanced NH3 volatilization and nitrification-denitrification but inhibited N fixation. The development of periphytic biofilms reduced NH3 volatilization loss but increased nitrification-denitrification loss and the overall N loss in the paddy system. The results from this work suggest that the presence of periphytic biofilms in paddy fields could increase N loss by 3.10%–7.11%. Deep fertilization is an effective method to retard the development of periphytic biofilms in the paddy system and can potentially increase the overall N use efficiency.     

黄土区长武塬土地利用变化对土壤水平衡的影响

[目的]揭示土地利用变化对土壤水平衡中各水文变量的影响,明确黄土区水循环演变机制,进而为相似研究提供技术支撑。[方法]以黄土区长武塬为研究区,采集农地、18 a和26 a苹果园下20 m土样,测定了土壤水及其稳定性和放射性同位素含量,分解量化了土壤水平衡中各分量,评估了其对土地利用变化的响应。[结果]以农地为基准,18 a和26 a苹果园土壤水储量减少7.3%～14.3%,深层渗漏量减少41.2%～80.0%,蒸散发量增加5.0%～9.6%。蒸散发中,蒸发量减小8.6%～17.2%,而蒸腾量增加了12.9%～25.4%。[结论]3种土地利用方式下土壤水含量存在显著差异,土壤水氧同位素剖面总体呈现先减小后稳定的变化趋势,农地条件下土壤水受到的蒸发最强。深根植物苹果树主要通过增加蒸腾耗水而减少其他水文变量的通量,显著影响水文过程。     

苹果树叶多病害及不可辨别病害的轻量识别算法

为提高苹果园中现有设备的病害树叶识别精度,该研究提出了一种快速识别方法S-DenseNet-E。首先,基于DenseNet的Dense模块提出了S-Dense模块,并基于S-Dense模块搭建了S-DenseNet模型。S-Dense模块在输出层中以前馈直连方式将模型内每一层输出聚合连接在一起,改善了Dense模块的密集连接存在计算量大的问题,有效减小了模型计算量。通过在Phytopathology 2021 FGVC8的苹果树叶病害公开数据集上测试表明,S-DenseNet的F1-score达到85.14%,高于常用的CNN类模型;其识别推理时间（或延迟）是33.03 ms,低于MobileNetV2模型。其次,针对S-DenseNet模型在不可辨别病害上的F1-score较低（65.82%）的问题。该研究在S-DenseNet基础上增加辅助模型专门识别不可辨别病害,形成S-DenseNet-E方法。通过在同一数据集上测试表明,S-DenseNet-E在不可辨别病害上的F1-score达到70.10%,识别推理时间为38.92 ms,比S-DenseNet模型仅升高5.89 ms,并且保持了原来S-DenseNet对其他病害的识别效果。因此,该研究表明,S-DenseNet-E方法针对苹果患多种病害和不可辨别病害两种复杂情形的识别效果好,并且计算资源的需求较少,满足果园实际需求。     

基于属性特征知识图谱的细粒度葡萄园害虫识别

针对葡萄园害虫识别过程中害虫种类繁多、形态属性复杂、实体间深层次关联关系差等因素导致的识别准确率不够高的问题,该研究提出一种基于属性特征知识图谱的细粒度葡萄害虫识别模型。该模型以视觉编码器作为提取图像高层表征的骨干网络,并结合知识图谱具有在描述害虫实体属性特征和害虫实体间关联方面的优势,将知识图谱所提供的细粒度属性特征和害虫实体关联特征信息用于葡萄园害虫分类研究。该方法在2个数据集上进行了测试：一是GP21数据集,该数据集由公开大规模害虫数据集IP102中21类葡萄园相关害虫类别图像构成;二是GP8数据集,该数据集包含从农业生产基地实地采集并手工标注的8种葡萄园害虫。试验结果表明,该文所提模型性能明显优于普遍通用预训练网络模型,分类准确率在GP21数据集和GP8数据集上分别达到了91.21%和95.03%,相比于仅使用视觉编码器分别增加1.64和1.17个百分点。这证明属性特征知识图谱的引入能够辅助视觉编码器捕获细粒度更高的葡萄园害虫特征信息,有效解决了葡萄园害虫识别中的精度不够高的问题。     

高效深栽造林钻孔机的研制与试验

荒漠化是当今世界威胁人类生存的重要问题,深栽造林技术可以有效提高干旱沙地造林成活率和保存率,所以高效深栽造林设备的研究就显得尤为重要。论文首先对国内外干旱沙地深栽造林方式及相关机具进行了综述比较,得出深栽造林钻孔机的造林成活率最高,但是造林效率还需要提高。在对工作环境进行了解的基础上,提出了深栽造林钻孔机的总体方案,以LH1630-2拖拉机为动力底盘,采用全液压驱动与控制的方式,完成了各个关键零部件的设计选型和样机的制作与装配,并在内蒙古林场进行了深栽造林钻孔实地试验。试验结果表明,深栽造林钻孔机钻孔过程比较顺畅,钻孔效率相比传统的单钻深栽造林钻孔机提高了约1倍,可达到180孔/h,钻孔直径为90 mm左右,钻孔最大深度可达1.7 m,钻孔间距可以在2.5~4 m之间调整,钻孔行距可以通过拖拉机的行驶在1~4 m之间调整,对于适合在本地区生长的杨树树种,深栽造林的成活率可达90%以上。研究结果可为深栽造林钻孔的设计和应用提供参考。     

Spectral signatures in the UV range can be combined with secondary plant metabolites by deep learning to characterize barley–powdery mildew interaction

In recent studies, the potential of hyperspectral sensors for the analysis of plant–pathogen interactions was expanded to the ultraviolet range (UV; 200–380 nm) to monitor stress processes in plants. A hyperspectral imaging set-up was established to highlight the influence of early plant–pathogen interactions on secondary plant metabolites. In this study, the plant–pathogen interactions of three different barley lines inoculated with Blumeria graminis f. sp. hordei (Bgh, powdery mildew) were investigated. One susceptible genotype (cv. Ingrid, wild type) and two resistant genotypes (Pallas 01, Mla1- and Mla12-based resistance and Pallas 22, mlo5-based resistance) were used. During the first 5 days after inoculation (dai) the plant reflectance patterns were recorded and plant metabolites relevant in host–pathogen interactions were studied in parallel. Hyperspectral measurements in the UV range revealed that a differentiation between barley genotypes inoculated with Bgh is possible, and distinct reflectance patterns were recorded for each genotype. The extracted and analysed pigments and flavonoids correlated with the spectral data recorded. A classification of noninoculated and inoculated samples with deep learning revealed that a high performance can be achieved with self-attention networks. The subsequent feature importance identified wavelengths as the most important for the classification, and these were linked to pigments and flavonoids. Hyperspectral imaging in the UV range allows the characterization of different resistance reactions, can be linked to changes in secondary plant metabolites, and has the advantage of being a non-invasive method. It therefore enables a greater understanding of plant reactions to biotic stress, as well as resistance reactions.     

深路堑边坡工程的设计理论及主要措施

随着我国西部开发力度的加大，大量公路、铁路、场地等工程也随之大量出现，涉及到大量的深路堑边坡，因此很多深路堑边坡设计问题也相应日益突出。本文简要阐述深路堑边坡工程的现状及发展情况、工程的实用性及设计原则、边坡变形失稳破坏类型及对应的设计处理措施。     

深层入渗对土壤含水率和苹果产量的影响

选取15年生开心形富士苹果,设计7个不同入渗方式处理,测定了不同处理的苹果产量和不同生育期土壤含水率。结果表明,苹果成熟前不同处理土壤含水率之间差异不明显,进入雨季以后,深层入渗技术能够将雨季降雨有效收集,并入渗到深层储存,在60～200cm,土壤含水率高出对照1%～5.22%;各处理之间苹果单株产量差异极显著,其中深层入渗技术都有增产效果,平均增产13.46%。深层入渗技术能够有效的使降雨入渗到更深层次并储存,为苹果需水临界期提供土壤水分保证,提高产量。