农业害虫检测的深度学习算法综述

害虫检测是害虫测报的关键步骤,对于害虫防治具有重要意义,也是保证农作物产量和品质的前提。近年来,随着卷积神经网络的迅速发展,害虫检测技术进入智能化时代,使用深度学习相关技术实现精确的害虫检测已成为研究人员重点关注的课题。为了促进深度学习害虫检测技术的发展,对检测算法和现有数据集进行综述。总结了当前面临的数据匮乏、小目标检测、多尺度检测和密集与遮挡检测等四大难点问题,并分析了其主要成因。重点针对以上难点问题,总结归纳了近年来提出的深度学习害虫检测算法的改进策略和技术细节,以及面向实际场景的应用算法,对比分析了各类算法的性能表现、改进策略的适用场景及其优缺点。从面向复杂检测场景、解决数据匮乏问题、模型增量更新和应用落地等方面分析并展望了未来的研究趋势。     

Domestic Water Consumption under Intermittent and Continuous Modes of Water Supply

Although an extensive literature emphasizes the disadvantages of intermittent water supply, it remains prevalent in rural areas of developing countries. Understanding the effects of water supply time restrictions on domestic water use activities and patterns, especially for hygienic purposes, is important for the elaboration of the water supply. We studied the influence of intermittent and continuous water supply on water consumption and related activities in villages in the central region of the Wei River basin, China. Data were collected from a survey of 225 households in the sampled villages. Compared with a continuous water supply of 24 h d^?1 (hours per day), adopting an intermittent water supply can reduce domestic water consumption. However, it presents risks in terms of hygiene behavior, particularly the frequency of face, hands, and feet washing, as well as water sharing among family members. Outdoor water consumption is more affected than indoor water consumption under slight supply restriction (≥6 and?<?24 h d^?1), whereas indoor water use is most affected under moderate supply restriction (>1.5 and?<?6 h d^?1). Villages with high supply restriction (≤1.5 h d^?1) meet only the minimum basic requirements for domestic use, 33.6–34.7 L c^?1 d^?1 (liters per capita per day). We conclude that the determination of the daily water delivery duration for intermittent water supply in rural communities of developing countries should give greater consideration to differences in water use activities and patterns under the water supply time restrictions.     

橡胶混凝土抗冲磨性能的研究

探讨了橡胶混凝土在不同水灰比、胶粒掺量、粒径大小的情况下的抗冲磨性能,为进一步改善混凝土的韧性和耐久性提供借鉴。在以往橡胶混凝土研究基础上,采用水下钢球法对混凝土进行抗冲磨强度试验,得出了橡胶混凝土在不同水灰比、胶粒掺量、粒径大小的情况下的抗冲磨性能。从本次试验中可以看出水灰比与橡胶混凝土的抗冲磨强度具有良好关系,水灰比小时,有利于提高混凝土的耐磨性,但是水灰比不能无限制的降低,必须能够提供胶凝材料完全水化所需的水分及保证施工过程中的和易性;随着橡胶颗粒掺量的增加,混凝土的耐磨性先提高后降低。橡胶掺量15%时混凝土的抗冲磨强度最高。当掺量达到30%时,抗冲磨强度反而比未掺入时低;掺入的三种粒径中掺入粒径为（2.36～4.75 mm）的混凝土耐磨性能最佳,其单位磨损量最小。胶粒掺量是影响抗冲磨性能的主要因素,其次为水灰比,第三是胶粒粒径的大小。     

基于GIS的区域坡度坡长因子提取算法

为提高基于地理信息系统的区域土壤侵蚀研究、水土保持环境效应评价、流域水文分析等的应用效率,设计新的坡度坡长(LS)因子算法,利用正向-反向遍历算法取代原累积坡长算法,以获取区域尺度下的LS因子。实验结果表明,在计算精度允许的范围内,新算法使计算机运行效率有较大幅度的提高。     

针对不确定正例和未标记学习的最近邻算法(英文)

研究了在正例和未标记样本场景下不确定样本的分类问题,提出了一种新的算法NNPU(nearest neighbor algorithm for positive and unlabeled learning)。NNPU具有两种实现方式:NNPUa和NNPUu。在UCI标准数据集上的实验结果表明,充分考虑数据不确定信息的NNPUu算法要比仅仅考虑样本中不确定信息均值的NNPUa算法具有更好的分类能力;同时,NNPU算法在对精确数据进行分类时,比NN-d、OCC以及aPUNB算法性能更优。     

Nonlinear continuous multi-response problems: a novel two-phase hybrid genetic based metaheuristic

Generally the most real world production systems are tackling several different responses and the problem is optimizing these responses concurrently. This study strives to present a new two-phase hybrid genetic based metaheuristic for optimizing nonlinear continuous multi-response problems. Premature convergence and getting stuck in local optima, which makes the algorithm time consuming, are common problems dealing with genetic algorithms (GAs). So we hybridize GA with a clustering approach and particle swarm optimization algorithm (PSO) to make a balanced relationship between time consuming and premature termination. The proposed algorithm also tries to find Ideal Points (IPs) for response functions. IPs are considered as improvement measures that determine when PSO should start. PSO based local search exploit Pareto archive solutions to enhance performance of the algorithm by expanding the search space. Since there is no standard benchmark in this field, we use two case studies from distinguished paper in multi-response optimization and compare the results with some of the mentioned algorithms in the literature. Results show the outperformance of the proposed algorithm than all of them.     

基因表达式编程的LFC方法及其应用

介绍了基因表达式编程的基本原理,提出了具有线性复杂度的个体适应度评估方法(LFC),并且在遗传操作中采用自适应代沟替代策略,解决了标准GEP算法在求解复杂问题时时空效率低的问题。仿真表明,与标准GEP算法相比,该算法在不损失解的质量的情况下,求解效率得到明显改善。     

Dynamic reliability analysis of spacial flexible beam

In order to investigate the dynamic response reliability of the spacial flexible beam, the Monomial Cubature Rules (MCR) based Stochastic Response Surface Method (SRSM) is applied to derive the surrogate model of the nonlinear implicit limit state function. A small set of sample pairs for flexible beam dynamic response is generated by taking the integral points of MCR as sample points. Based on these sample pairs and SRSM, the surrogate model of the implicit limit state function is established. The results show that the proposed method has an excellent performance in terms of accuracy and efficiency compared with the Probabilistic Collection Method (PCM) based SRSM and Monte Carlo method, and that the dispersion of system parameters has a significant effect on the dynamic response reliability of the spacial flexible beam.     

基于集成学习的N6甲基化位点预测方法研究

N6-甲基腺嘌呤(N6-methyladenine,6mA)是指腺嘌呤第6位氮原子的甲基化修饰。6mA在维持细胞正常的转录活性、DNA损伤修复、染色质重塑、遗传印记、胚胎发育和肿瘤发生等生物过程中起着非常重要的作用。通过生物实验的方法来鉴定6mA位点耗时且昂贵。近年来,研究界提出了一些基于机器学习的6mA位点预测方法,但这些预测方法过度依赖一种学习模型,导致模型的泛化能力不足以及预测的准确度不高等问题。集成学习综合多种预测模型的优点,具有较好的泛化能力及预测性能。因此,为了进一步提升6mA位点的预测准确性,提出了一种基于stacking集成学习的6mA位点预测模型Stack6mAPred。该模型由两层分类器组成,第一层集成了朴素贝叶斯、支持向量机(support vector machine,SVM)和LightGBM等三种主流分类器,第二层使用逻辑回归(logistic regression,LR)分类器。Stack6mAPred利用增强核苷酸组成等5种特征对实验已鉴定6mA序列和非6mA序列进行编码,使用XGBoost(extreme gradient boosting)算法进行特征选择,去除冗余特征。通过在水稻基准数据集上进行五折交叉验证,与目前性能最优的方法MM-6mAPred相比,Stack6mAPred在敏感性、特异性、准确度、MCC和AUC上均获得了更好的性能,分别提高了1.7%、1.36%、1.72%、0.06和0.031。     

CFRP在水工钢闸门修复中的应用研究

为解决水利工程中常用水工钢闸门的锈蚀问题,针对粘钢加固技术存在的闸门加固部位更易锈蚀、加固部位易产生应力集中等问题,充分发挥纤维增强复合材料(FRP)比强度高、耐腐蚀性好、不改变母材特性等优点,提出采用CFRP修复的方法。首先采用与试验结果对比的方式验证数值方法分析CFRP修复的可行性和正确性,然后采用数值方法对某水工钢闸门在未锈蚀、锈蚀和CFRP修复三种状态下的强度和刚度进行分析。分别采用改变单元厚度法和共结点重单元法对锈蚀和CFRP修复的闸门模型进行数值模拟。结果表明:以未锈蚀闸门的计算结果为基值,主横梁腹板、支臂腹板和面板中下部发生2 mm锈蚀时,主梁、支臂、面板的强度和刚度分别降低25.95%和23.73%、25.52%和24.24%、24.43%和25.00%。采用单层厚0.167 mm的CFRP修复后,主梁、支臂、面板的强度和刚度分别提高30.28%和26.43%、29.46%和24.53%、28.47%和29.87%。采用CFRP修复加固水工钢闸门具有明显的效果,研究成果可为闸门规范的修订完善和实际工程中采用CFRP修复提供理论依据。