基于改进YOLOv5s的CNN-Swin Transformer森林野生动物图像目标检测算法

【目的】为提高野生动物在复杂森林环境中的检测精度,促进森林野生动物保护技术发展,提出一种基于YOLOv5s网络模型、针对陷阱相机所摄取森林野生动物图像的改进检测算法。【方法】以包含湖南壶瓶山国家级自然保护区几种典型森林野生动物在内的数据集为研究对象,首先,对真实标注框图像进行裁剪、归一化和缩放处理,随机将2~4张裁剪图像拼贴组成新的数据集元素,以丰富和增强数据集图像信息;其次,使用一种基于通道注意力思想的加权通道拼接方法,在通道拼接时引入权重改变通道数量,通过反向传播训练方法不断更新权重以增加重要特征信息的通道层数;接着,引入Swin Transformer模块与CNN网络相结合,为卷积神经网络特征提取加入自注意力机制,融合2种网络特征提取层的优势,提高特征提取的感受野;最后,选择更优的α-DIoU损失函数替代GIoU损失函数,针对边界框重叠面积和中心点距离造成的损失,引入新的几何因素惩罚项。【结果】在相同试验条件和数据集下,相比原YOLOv5s网络模型,改进算法极大提高检测的平均准确率和平均回归率,均值平均精度由74.1%提升至88.4%,获得14.3%的精度提升,同时也超过YOLOv3、YOLOXs、RetinaNet、Faster R-CNN等其他流行目标检测算法。【结论】针对陷阱相机所摄取森林野生动物图像背景与目标对比度低、遮挡重叠严重,致使检测误检率、漏检率高等问题,在检测算法中提出一系列改进措施,为我国森林野生动物的保护和数据获取提供一种新的可行性方案和思路。

Gender and rural transformation:A systematic literature review

Rural transformation can improve poverty reduction,living standards,and health outcomes in developing countries.However,impacts associated with rural transformation vary by region,household,and individual trait(including gender).While research on rural transformation has been increasing over the last decade,there has been no comprehensive review conducted on the relationships between gender and rural transformation.Here,we conduct a systematic literature review to investigate the impacts of rural transformation on gender and the influence of gender inclusiveness on rural transformation.We reviewed 82 studies from 1960-2021 that explore the relationships between rural transformation and gender.We then developed a framework that captures incidences and flow directions between indicators.Results show that most studies examined the impacts of rural transformation on women and between gender indicators.Few investigated the role of women and the influence of gender inclusiveness on rural transformation.Overall,studies showed that rural transformation typically leads to positive outcomes for women regarding employment,income,and empowerment.However,negative impacts on women’s control over income,stability of new income sources,and access to healthy food are also common.Tailoring future development policies and programs to explicitly account for gender inclusiveness can lead to more successful rural transformation.

Teachers' Transformation and the Way to Realize the Construction of Entrepreneurial Universities

The fundamental problem of transforming traditional universities into entrepreneurship universities lies in the transformation of teachers. The development of entrepreneurial universities has an important impact on the global higher education system, especially on the role of University teachers. Therefore, the author carries out research on the path of teachers' transformation and the construction of entrepreneurial universities. The study found that the traditional university teacher role has emerged as a new feature. How to realize the transformation of the entrepreneurial university teacher role, ease the identity crisis of the teacher role, and make it better adapt to the professional development and the construction of the entrepreneurial university. An important research topic. In order to realize the transformation of traditional teachers to entrepreneurial teachers, it is necessary to change the employment standards and establish a school-level achievement transformation platform.

Teachers'Transformation and the Way to Realize the Construction of Entrepreneurial Universities

The fundamental problem of transforming traditional universities into entrepreneurship universities lies in the transformation of teachers.The development of entrepreneurial universities has an important impact on the global higher education system,especially on the role of University teachers.Therefore,the author carries out research on the path of teachers'transformation and the construction of entrepreneurial universities.The study found that the traditional university teacher role has emerged as a new feature.How to realize the transformation of the entrepreneurial university teacher role,ease the identity crisis of the teacher role,and make it better adapt to the professional development and the construction of the entrepreneurial university.An important research topic.In order to realize the transformation of traditional teachers to entrepreneurial teachers,it is necessary to change the employment standards and establish a school-level achievement transformation platform.

Research on the Transformation of Teachers'Functions in New Applied Colleges

With the continuous development of society,the demand for applied talents is becoming more and more urgent,and the new applied undergraduate colleges based on the cultivation of applied talents can effectively solve this contradiction between supply and demand.If we want to realize the cultivation of applied talents,we need to establish a team of teachers with excellent quality.Therefore,this paper gives four ways to solve this problem:strengthening the construction of teachers’ethics,improving teachers’ideological and political ability,strengthening teachers’curriculum practice level,and improving teachers’ability of selfcareer planning.

Research on the Transformation and Development of College English Teachers under the Mode of Applied Talents Training

With the rapid development of economic globalization and the transformation of higher education from elite education to popular education,all employers require college students’English ability is getting higher and higher,especially for the ability to work in English.Teachers not only play a major role in the process of training students’application ability,but also determine the success or failure of the transformation and development of local colleges and universities.This also puts forward higher requirements for College English teachers.The majority of English teachers must realize their own transformation and development from the overall situation of the transformation and development of colleges and the actual needs of students to learn English.College English teachers should adapt to the new situation,change teaching ideas and establish the concept of lifelong learning,so as to increase their own practical experience and industry experience,so as to promote the development of College English teaching.

Research on the Transformation of Teaching and Research Form of Professional Teachers in Blended Learning at Colleges and Universities - Taking the Java Programming Course as an Example

In view of the current situation that offline teaching is the main mode of teaching Java Programming in higher vocational schools,this paper introduces the online and offline hybrid teaching method and expounds it from the aspects of blended learning design,teaching organization,and implementation.At the same time,combined with the characteristics of blended learning,this paper proposes that under the new mode,teachers should actively change the form of teaching and research,the teaching mode,and the role of teachers,take students as the center,and build an independent and effective classroom.

基于S-YOLO V5和Vision Transformer的视频内容描述算法

视频内容描述的自动生成是结合计算机视觉和自然语言处理等相关技术提出的一种新型交叉学习任务。针对当前视频内容生成描述模型可读性不佳的问题,本研究提出一种基于S-YOLO V5和Vison Transformer(ViT)的视频内容描述算法。首先,基于神经网络模型KATNA提取关键帧,以最少帧数进行模型训练;其次,利用S-YOLO V5模型提取视频帧中的语义信息,并结合预训练ResNet101模型和预训练C3D模型提取视频静态视觉特征和动态视觉特征,并对两种模态特征进行融合;然后,基于ViT结构的强大长距离编码能力,构建模型编码器对融合特征进行长距离依赖编码;最后,将编码器的输出作为LSTM解码器的输入,依次输出预测词,生成最终的自然语言描述。通过在MSR-VTT数据集上进行测试,本研究模型的BLEU-4、METEOR、ROUGEL和CIDEr分别为42.9、28.8、62.4和51.4;在MSVD数据集上进行测试,本研究模型的BLEU-4、METEOR、ROUGEL和CIDEr分别为56.8、37.6、74.5以及98.5。与当前主流模型相比,本研究模型在多项评价指标上表现优异。

基于改进YOLOv5s 的白酒瓶盖瑕疵检测

目的瓶装白酒生产过程中,瓶盖表面瑕疵会影响产品外观质量。针对白酒瓶盖表面瑕疵检测效率低和目标检测效果差的问题,提出一种基于YOLOv5s的改进算法DTS-YOLO。方法首先,在主干网络中引入可变形卷积,以提高模型对极端长宽比瑕疵的检测精度。其次,引入Transformer编码块,使网络聚焦于提取图像的全局信息。最后,在颈部网络构建C3SE-Lite模块,将C3模块嵌入SE注意力模块的同时引入Ghost卷积,减少参数量的同时,增强对瓶盖瑕疵的检测能力。结果实验结果表明,本文所提方法相较于基础网络,参数量减少了10%,平均精度均值达95%,平均检测速度达30帧/s。结论本文方法有效实现了白酒瓶盖表面瑕疵快速、准确地检测,可广泛应用于瓶装白酒生产过程中瓶盖表面检测。

Sora变革教育:基于教师角色与学习方式的洞见

生成式人工智能的发展又迎来重大飞跃,具有“通用物理世界模拟器”潜力的文生视频模型——Sora被推出。技术的革新势必对教育领域产生重大影响,Sora对教师角色的影响如何?又将为学习方式带来哪些变革?回答好这两个问题是推动Sora变革与赋能教育领域的关键。基于此,本研究在对Sora进行本质解析的基础上,分析Sora对教师角色产生的影响,包括Sora对教师角色挑战的冲击、Sora变革下教师何为的应对以及Sora与教师协同发展的共进;接着提出Sora对学习方式的变革,包括进行“可想即可见”的学习、虚实融合学习的升级、人机协同学习的革新、促进创造性学习实现以及丰富个性化学习形态等;最后对Sora变革教育进行审思与展望,包括Sora的教育边界确立、Sora的使用标准建立、Sora应用的算力需求以及开发教育专属Sora模型。研究以期促进Sora有效应用于教育领域,提升学习效果与质量。

改进的YOLOv5s遥感影像机场场面飞机小目标识别

遥感影像具有地物拍摄模糊以及背景环境复杂等特点,导致大面积地面物体识别准确率较低。针对此问题,提出了一种基于YOLOv5s改进网络模型。该模型对YOLOv5s中的骨干提取网络和颈部多尺度特征融合网络进行优化,引入Swin Transformer以获得更多关于目标物体的特征信息;同时对骨干网络中的模块做了修剪;此外,模型中还添加了坐标注意力机制来提升特征提取和融合效果。对于遥感数据集进行了小目标检测识别测试实验,改进后的YOLOv5s网络的mAP值为0.8375,比YOLOv5s网络模型提高了0.0225。实验结果表明,改进后的YOLOv5s网络模型对比YOLO系列网络和EfficientDet模型有效地提高了识别准确率、召回率以及mAP值,并且在训练时间上也比YOLOv5s减少了1/12。

融合Transformer和改进PANet的YOLOv5s交通标志检测

针对交通标志检测速度慢和目标大小与类别极度不平衡等问题,提出一种融合Transformer和改进PANet网络的YOLOv5s交通标志检测算法。首先在不增加模型复杂度的前提下,将主干网络末端与Transformer融合以提高网络特征提取能力;其次由于所采用交通标志数据集的目标尺度太小,导致网络32倍大尺度检测层检测效果不佳,故不采用相关网络层,同时采用K-means算法得出适合的预测候选框;然后改进损失函数以解决正负样本极度不平衡问题。最后将所提出的改进算法在Jetson AGX Xavier平台上部署验证。实验结果表明,所提算法检测性能更佳,其准确率和召回率在原网络的基础上分别提高了2.2%和0.7%,模型参数量和计算复杂度分别减少了25.8%和10.1%。在Xavier上的检测速度达到76FPS,满足实时交通标志检测的要求且易于在实际场景部署。

改进YOLOv5s的钢材表面缺陷检测算法

为提高钢材表面缺陷检测的准确率,提出一种改进YOLOv5s的钢材表面缺陷检测算法。首先,在特征提取网络中引入Swin Transformer结构,增强网络对特征的感知能力;其次,添加坐标注意力机制,加强对重要特征信息的关注;最后,针对钢材缺陷的特点增加检测层,提升多尺度目标检测能力,并使用SIOU损失函数评估检测效果。将所提出的算法在公开数据集NEU-DET上进行消融实验,结果表明:所提算法能有效提高钢材表面缺陷目标检测的准确率。

K-S变换及其电网超谐波时频分析应用

依据FFT→优化窗→IFFT思路,突破线性时频变换的窗函数积分性能桎梏,实现高性能优化窗函数的线性时频变换应用,建立新型时频变换算法——K-S变换.对信号x(t)的FFT频谱向量进行频移处理后,与该频移点下Kaiser优化窗的频谱向量进行Hadamard乘积,再将乘积结果进行FFT逆变换(IFFT),构造出K-S变换复时频矩阵,由此获得x(t)的时间-频率-幅值、时间-频率-相位三维信息;给出逆变换的数学推导与局部性质、线性性质和变分辨率特性;0~150 kHz电网的稳态与时变超谐波信号仿真实验表明,K-S变换的时域、频域分辨能力均优于流行的短时傅里叶变换、S变换,具有优良的变分辨率性能;0~40 kHz超谐波信号的实测证明,基于K-S变换的超谐波电压幅值测量绝对误差均小于0.032 3 V.

基于哨兵函数和S变换的风力机叶片材料损伤特性研究

利用声发射检测技术研究玻璃纤维增强环氧树脂复合材料的损伤特性,在此过程中,采用哨兵函数来表征该材料的损伤程度,并通过S变换和模糊C均值(FCM)聚类来分析声发射信号,从而获得材料的损伤特征。三点弯曲实验结束后对试件断口进行扫描电子显微镜(SEM)拍照来验证,可得:通过对SEM照片的分析得到基体开裂、纤维脱粘、分层破坏、纤维断裂4种损伤模式;对整个声发射事件进行哨兵函数分析,观察到试件在弯曲过程中哨兵函数曲线呈明显下降趋势;对依据哨兵函数划分的不同阶段的信号进行VMD降噪处理,然后采用S变换进行时频分析得到不同损伤的特征频率,最后采用FCM聚类进行验证。结果表明:哨兵函数值的突变可作为材料断裂的预警信号,材料损伤类型的识别可依据S变换的频率分布结果进行确定。

基于S变换双阈值法的汽车零部件载荷谱加速编辑

针对S变换编辑法加速编辑效果不佳的问题,提出了一种基于S变换双阈值编辑方法。该方法先对载荷谱进行S变换以获取最大幅值谱,在以双阈值识别并保留的幅值谱片段为依据保留对应的载荷谱片段后,再将其拼接成加速后的载荷谱,最后对比分析S变换双阈值编辑法与S变换编辑法编辑的加速谱的统计参数、功率谱密度、穿级计数和疲劳仿真结果。研究结果表明,S变换双阈值编辑法可明显压缩原始载荷时间,且其压缩效率高于S变换编辑法,转向节的疲劳损伤和寿命分析误差更小,适用于汽车零部件载荷谱加速编辑研究。

从协同到求异:指向教师专业发展的U-S合作优化研究

U-S合作是教师专业发展的重要举措。审视U-S合作的历程可知,U-S合作致力于多方位促进教师专业发展,协同逐渐成为U-S合作促进教师专业发展的底层逻辑。随着合作的持续推进,这一底层逻辑在促进教师专业发展方面所潜在的矛盾也逐渐暴露出来。一方面,大学与中小学的文化差异阻碍了协同的深度发生;另一方面,为了实现大学与中小学的协同,双方不得不做出诸多的妥协,在妥协的语境下,中小学教师往往面临着主体性丧失的风险。与之相对,在求异这一核心理念的引领下,跨界学习旗帜鲜明地提出了“差异即资源”的主张,并通过互动场域的营造和互动系统的建构,为“差异即资源”提供了现实可能,也为优化U-S合作提供了新的思路。

基于S变换模能量分析的直流电缆故障测距

针对现有高压直流电缆故障定位方法故障定位精度低、定位模型泛化能力差等问题,文中在考虑电缆故障后故障点产生的暂态信号中含有丰富的故障信息的特征,提出一种时频域能量分析的故障测距方法。该方法利用多分类支持向量机(SVM)的特性构建直流电缆故障定位模型,利用直流电缆相模解耦矩阵对电缆电气量解耦,并采用S变换进行时频域变换,将得到的模量构造模能量,将其作为样本对SVM模型训练。实验结果表明,所提方法对于不同故障类型在不同过渡电阻及不同位置故障下均可进行准确测距。

S-100β CysC和NF-κB对急性缺血性脑卒中患者静脉溶栓后出血转化的预测价值

目的 探讨急性缺血性脑卒中(AIS)外周血中S-100β、CysC和NF-κB水平对静脉溶栓后出血转化的影响及预测价值。方法 收集2019-03—2022-03接受溶栓治疗的AIS患者140例,根据溶栓后24 h是否发生出血转化(HT)将患者分为非HT组(n=112)和HT组(n=28)。比较2组一般临床资料,采用多因Logistic回归分析影响HT发生的危险因素,受试者工作特征(ROC)曲线评估S-100β、CysC和NF-κB预测HT发生的价值。结果 HT组与非HT组相比,患者年龄、发病至溶栓时间、房颤、TOAST分型、C反应蛋白、凝血酶原时间、S-100β、CysC、NF-κB、白质高信号和脑微出血等均有统计学差异(P<0.05)。Logistic多因素回归分析显示,房颤、S-100β、CysC和NF-κB为影响HT发生的危险因素。S-100β、CysC和NF-κB预测AIS患者静脉溶栓后出血转化的曲线下面积分别为0.915(0.902~0.923)、0.874(0.856~0.882)和0.789(0.771~0.796),均具有一定的预测价值。结论 S-100β、CysC和NF-κB为AIS患者静脉溶栓后HT发生的危险因素,对HT发生具有一定的预测价值。