基于改进YOLOv8n的茶叶嫩稍检测方法

针对名优茶智能采摘中茶叶嫩梢识别精度不足的问题,该研究对YOLOv8n模型进行优化。首先,在主干网络中引入动态蛇形卷积（dynamic snake convolution,DSConv）,增强模型对茶叶嫩梢形状信息的捕捉能力;其次,将颈部的路径聚合网络（path aggregation network,PANet）替换为加权双向特征金字塔网络（bi-directional feature pyramid network,BiFPN）,强化模型的特征融合效能;最后,在颈部网络的每个C2F模块后增设了无参注意力模块（simple attention module,SimAM）,提升模型对茶叶嫩梢的识别关注度。试验结果表明,改进后的模型比原始模型的精确率（precision,P）、召回率（recall,R）、平均精确率均值（mean average precision,mAP）、F1得分（F1 score,F1）分别提升了4.2、2.9、3.7和3.3个百分点,推理速度为42 帧/s,模型大小为6.7 MB,满足低算力移动设备的部署条件。与Faster-RCNN、YOLOv5n、YOLOv7n和YOLOv8n目标检测算法相比,该研究提出的改进模型精确率分别高出57.4、4.4、4.7和4.2个百分点,召回率分别高出53.0、3.6、2.8和2.9个百分点,平均精确率均值分别高出58.9、5.0、4.6和3.7个百分点,F1得分分别高出了56.8、3.9、3.7和3.3个百分点,在茶叶嫩梢检测任务中展现出了更高的精确度和更低的漏检率,能够为名优茶的智能采摘提供算法参考。     

基于含水率变化的气体射流冲击干燥过程温度自适应控制系统

为了给变温干燥工艺提供新的技术支持,实现基于含水率变化的干燥温度自适应控制,该研究设计了具有物料含水率在线检测功能的温度自适应控制系统。采用卷积神经网络建立了以质量检测值、气流冲击速度、称重传感器弹性基体温度、气流冲击距离为输入,物料真实质量为输出的含水率在线检测模型。进行了含水率在线检测模型验证试验。结果表明,该模型满足变温干燥工艺中含水率在线检测的精度要求,5组含水率在线检测模型验证试验的决定系数R²和均方根误差RMSE依次为0.9934和1.20%。该文设计了改进神经网络-PID（improved neural network-PID,INN-PID）控制器来实现变温干燥工艺中的温度控制。在MATLAB软件中以单位阶跃信号为输入对PID、神经网络-PID（neural network-PID,NN-PID）和INN-PID控制器的动态性能进行仿真。对3种控制器分别进行了50~55 ℃的干燥温度控制试验。结果表明,在仿真试验中,INN-PID控制器的控制稳定性和调节时间均显著优于另外两种控制器;干燥温度控制试验结果与仿真结果存在近似相同的规律,INN-PID控制器的峰值时间是208.00 s,调节时间是120.59 s,最大超调量是4.87 %,满足变温干燥过程中温度控制的要求。该研究在气体射流冲击干燥机中搭建了温度自适应控制系统,进行了基于含水率变化的温度自适应控制试验。结果表明,该系统可以对基于含水率变化的变温干燥工艺中的干燥温度进行快速且有效的调节。该研究对提高干燥设备的自动化水平以及开发新的变温干燥工艺具有重要意义,对其他领域的多信息融合检测和控制策略研究提供参考。     

水下底栖生物轻量化目标检测算法YOLOv7-RFPCW

针对水下底栖生物检测模型参数量过多,计算量过大,同时水下环境复杂容易造成错检和漏检,影响模型在水下底栖生物检测任务中的准确性的问题。提出了一种水下底栖生物轻量化检测算法YOLOv7-RFPCW。对YOLOv7网络重新设计轻量级网络结构,降低了特征提取网络的参数量和计算量,减少模型体积。设计了P-ELAN和P-ELAN-W模块,进一步轻量化特征提取网络;针对水下图像颜色失真,目标的空间位置不准确的问题,采用CBAM注意力模块加强特征融合,减少信息丢失,以更好地适应特殊的水下环境;针对水下目标容易出现形状变形的问题,使用WIOU-V3损失函数替换默认的CIOU损失函数,提高水下底栖生物检测的鲁棒性。试验结果显示,改进后的模型YOLOv7-RFPCW的参数量和计算量分别减少了75.9%和30.7%,模型体积减小了75.3%,精度提升了1.9个百分点。这一综合性的提升兼顾了轻量化和精度,为在水下环境中部署提供了可行的解决方案,验证了所提出的改进算法能胜任水下底栖生物检测任务。     

基于改进YOLOv7的杨梅树单木检测

为了快速检测和统计杨梅树的数量,该研究提出了一种基于改进YOLOv7的杨梅树单木检测模型：YOLOv7-ACGDmix。首先,对YOLOv7的可扩展高效长程注意力网络（extended-efficient long-range attention networks, E-ELAN）进行改进,通过融合兼具卷积和注意力机制优势的ACmix（a mixed model that enjoys the benefit of both self-attention and convolution）结构得到AC-E-ELAN模块,提升模型的学习和推理能力,引入可变形卷积（deformable convolutional networks version 2, DCNv2）结构得到DCNv2-E-ELAN模块,增强模型对不同尺寸目标的提取能力;其次,采用内容感知特征重组（content-aware reassembly of features, CARAFE）上采样模块,提高模型对重要特征的提取能力;然后,在主干和头部网络部分添加全局注意力机制（global-attention mechanism, GAM）,强化特征中的语义信息和位置信息,提高模型特征融合能力;最后,采用WIoU（wise intersection over union）损失函数减少因正负样本数据不平衡造成的干扰,增强模型的泛化性。在公开数据集上的试验结果表明,YOLOv7-ACGDmix模型的精确率达到89.1%,召回率达到89.0%,平均精度均值（mean average precision, mAP）达到95.1%,F₁-score达到89.0%,相比于原YOLOv7模型分别提高1.8、4.0、2.3和3.0个百分点。与Faster R-CNN、SSD、YOLOv8模型相比,改进模型的平均精度均值（mAP_0.5）分别提高了9.8、2.2、0.7个百分点。实地采集杨梅树样本数据的检测精确率87.3%、召回率85.7%。试验表明,改进模型为基于无人机影像的杨梅树单木检测提供了一种有效的解决方案,对果园精准管理的发展具有重要意义。     

改进YOLOv5s对病理学图像中猪只小肠绒毛的检测

为解决传统小肠绒毛需要专业人员手动检测耗时耗力且存在主观性和不稳定性等问题,同时提高在复杂病理学图像中小肠绒毛检测的准确率和效率,该研究提出基于改进YOLOv5s检测复杂病理学图像下猪只小肠绒毛的方法。首先,采用串联形式的混合池化对空间金字塔进行优化,增强特征提取与特征表达,提升检测精度;然后引入一种基于注意力机制的网络模块(simpleattentionmechanism,SimAM)与Bottleneck中的残差连接相结合,使用SimAM对Bottleneck中的特征图进行加权,得到加权后的特征表示,利用注意力机制加强模型对目标的感知。试验结果表明,该研究算法的平均精度（average precision）和每秒传输帧数（frame per second,FPS）达到92.43%和40帧/s。改进后的YOLOv5s在召回率和平均精度上相较改进前提高2.49和4.62个百分点,在不增加模型参数量的情况下,每帧图片的推理时间缩短1.04 ms。与经典的目标检测算法SSD、Faster R-CNN、YOLOv6s、YOLOX相比,平均精度分别提高15.16、10.56、2.03和4.07...     

PCR-based detection of Xanthomonas campestris pathovars in Brassica seed

Xanthomonas campestris is a seedborne bacterium that causes black rot of crucifers. Substantial crop losses may result from the rapid spread of the bacteria under favourable conditions, especially those occurring during seedling production. A PCR-based method has been developed for the rapid and sensitive detection of the pathovars of X. campestris that affect crucifers. Primers were designed to specifically amplify a 619 bp fragment of the hrpF gene from X. campestris . Amplification products were not detected from other Xanthomonas species, or from other pathogenic or epiphytic bacteria occurring on these plants. To avoid false-negative results arising from the presence of amplification inhibitors in plant extracts, primers targeting a 360 bp section of the internal transcribed spacer (ITS) region from Brassica spp. were included in a multiplex PCR. The assay readily detected X. campestris pv. campestris infections in diseased plants and from bacterial colonies isolated on growth media, and was more sensitive and specific than traditional plating methods and a commercially available ELISA. A seed-washing protocol was optimized to allow the detection of a single artificially infected seed among 10 000 healthy seeds using the multiplex PCR.     

心水病病原体高度保守基因合成与原核表达

通过软件DNAStar分析,设计合成心水病病原体抗原性、保守性比较高的一段基因,长度为564bp,将它插入pUC57载体,再转入pET-32a质粒中进行原核表达并纯化表达产物。结果表明,合成的基因在大肠杆菌中得到高效表达并能简易地提纯,纯化的表达产物有望用于心水病免疫学检测;此合成外来基因还可以作为心水病病原PCR检测的阳性对照。这些工作为我国建立心水病防控技术储备奠定了一些基础,也为防控高度危险性外来传染病提供一种安全和可行的研究新思路。     

利用巢式PCR快速鉴定牛传染性鼻气管炎

为了快速检测牛传染性鼻气管炎,采用SDS-蛋白酶K法,提取病毒模板DNA。根据IBRV gB基因序列设计了1对特异引物,在其上下游引物的内侧又分别设计了1对引物,以这4条引物对IBRV模板进行扩增。结果成功扩增出预期目的片段,建立了巢式PCR检测方法。敏感性、特异性等检测试验结果表明该方法能特异检测IBR病毒。本方法具有快速、灵敏、特异的优点,适用于在牛及其遗传物质的进出口检疫中进行牛传染性鼻气管炎快速病原鉴定。     

Stand dynamics and tree quality response to precommercial thinning in a northern hardwood forest of the Acadian forest region: 23 years of intermediate results

In the late 1980s, large forest companies began precommercial thinning (PCT) operations in young northern hardwood cutovers in New Brunswick, Canada. To provide supporting growth and yield information, an industrial experiment was established at residual stand densities of 1300, 1600, 1900, and 2200?stems?ha^?1. Stand responses were examined for measurements recorded at 0 (1987), 5 (1992), 10 (1997), 16 (2003), and 23 (2010) years after establishment. Average diameter at breast height, quadratic mean diameter, stand basal area, and stand total volume growth increased as stem density decreased from PCT. There were significant linear differences for many of these variables between treatments and time periods (year). No significant differences were detected in tree height between treatments. In 2010, the four PCT thinning treatments did not exhibit any differences in potential sawlogs at 2.4?m (8?ft) and 3.6?m (12?ft) lengths. Significant differences were observed for 4.9?m (16?ft) sawlogs that were produced at the least dense spacing (1300?stems?ha^?1). Results from this study and recommendations from the European literature suggest that value-added timber products may be produced from more intense PCT treatments than are currently being practiced on sites dominated by yellow birch (Betula alleghaniensis Britt.).     

STABILIZATION OF DUNE SAND AT BORGO BONSIGNORE IN SICILY / STABILISASIE VAN DUINSAND BY BORGO BONSIGNORE IN SICILIE

The methods employed and the results achieved in stabilizing an area of about 225 ha dune sand is described. The techniques employed in allowing tree species to become established are mentioned and are illustrated by various plates