基于全U网络的混凝土路面裂缝检测算法

针对现有的混凝土裂缝检测算法在各种复杂环境中检测精度不够、鲁棒性不强的问题,根据深度学习理论和U-net模型,提出一种全U型网络的裂缝检测算法。首先,依照原U-net模型路线构建网络;然后,在每个池化层后都进行一次上采样,恢复其在池化层之前的特征图规格,并将其与池化之前的卷积层进行融合,将融合之后的特征图作为新的融合层与原U-net网络上采样之后的网络层进行融合;最后,为了验证算法的有效性,在测试集中进行实验。结果表明,所提算法的平均精确率可达到83.48%,召回率为85.08%,F1为84.11%,相较于原U-net分别提升了1.48%,4.68%和3.29%,在复杂环境中也能提取完整裂缝,保证了裂缝检测的鲁棒性。     

Object-oriented DBMS requirements

The requirements on an object-oriented DBMS for management of information in a large, complex enterprise are presented. These requirements aid in the achievement of an environment characterized by data sharing, open architectures, application and data portability, and assurance of data integrity. They were defined from the point of view of a user of the DBMS; therefore they describe the expected functionality of the DBMS and do not specify the method of implementation to achieve this functionality. They encompass requirements on the data model, query and data manipulation languages, the system architecure, interfaces to the system, change management, and transaction management.     

Data Squashing by Empirical Likelihood

Data squashing was introduced by W. DuMouchel, C. Volinsky, T. Johnson, C. Cortes, and D. Pregibon, in Proceedings of the 5th International Conference on KDD (1999). The idea is to scale data sets down to smaller representative samples instead of scaling up algorithms to very large data sets. They report success in learning model coefficients on squashed data. This paper presents a form of data squashing based on empirical likelihood. This method reweights a random sample of data to match certain expected values to the population. The computation required is a relatively easy convex optimization. There is also a theoretical basis to predict when it will and won't produce large gains. In a credit scoring example, empirical likelihood weighting also accelerates the rate at which coefficients are learned. We also investigate the extent to which these benefits translate into improved accuracy, and consider reweighting in conjunction with boosted decision trees.     

一种3V CMOS音乐集成电路的设计原理

介绍了一种数字式全定制音乐集成电路设计原理。对伪音阶发生器、节拍发生器等子电路作了较详细的叙述。采用这种原理可设计出单芯片驱动压电陶瓷和发光二极管的音乐集成电路，该电路无须外接电阻、电容，可存储数首歌曲；另外，还可设计出一种数字音频程控频率合成器，时钟为ｆ０时，可控频率范围为ｆ０２～ｆ０２Ｎ。     

SOI全介质隔离与高频互补双极兼容工艺

ＳＯＩ材料的全介质隔离技术与高频互补双极工艺的结合是研制抗辐照能力强、频带宽、速度高的集成运算放大器的理想途径，从实验的角度提出了一种ＳＯＩ材料全介质隔离与高频互补双极工艺兼容的工艺途径。     

Full Or-Parallemism and Restricted And-Parallelism in BTM

BTM is a new And/Or parallel execution model for logic programs which exploits both full Or-parallelism and restricted And-parallelism.The advantages of high parallelism and low run time cost make BTJ,an experimental execution system of BTM implemented on a nonshared-memory multiprocessor system,achieve significant speedup for both And-paralled and Or-parallel logic programs.     

Sensitivity of generation reserve requirements in deregulated power systems

Since demand for power exhibits great variability the amount of firm energy to be purchased to meet real time demand based on forecasts is usually different from the realized demand for that period. The role of generation reserves is to meet the real time fluctuations of power demand. The predictable part of the demand is met through purchases of firm energy. In this paper a model is presented to determine optimal quantities of firm energy and generation reserves to meet random demands. The model is then parameterized introducing a set of factors to perform a sensitivity study. A full factorial experiment is designed to study the impact of five factors on the response variable (i.e., proportion of generation reserves on the total purchased quantity). An example consisting of 640 simulations corresponding to 2⁵ treatment combinations evaluated over 20 randomly generated mean demands is used to identify significant factors on the response variable. Results from the experiment suggest that generation reserve requirements should be adjusted considering changes in significant factors and in the mean demand over the dispatch interval.     

纯净钢生产技术及现状

本文简述了近年来国外主要炼钢厂生产纯净钢钢中杂质元素控制技术的发展，系统分析了钢中氧、碳、氮、硫、磷、氢等元素在钢中行为和降低杂质元素的措施。     

多重网格扩散综合加速的三维离散纵标中子输运程序SN3C

为提高已开发三维中子输运数值计算程序SN3C的效率,本文采用一致的扩散综合加速方法加速源迭代,并用Chebyshev加速方法提高幂迭代的收敛速度.针对扩散综合加速方法中求解扩散方程的特点,发展了基于完全多重网格方法的求解技术,数值试验的结果显示,SN3C的计算结果与Monte Carlo方法以及EVENT的计算结果符合很好;并且,对于中小规模的问题,完全多重网格方法的效率可与具有最优松弛因子的块超松弛方法相比;对于较大规模的问题,完全多重网格方法的性能更好.     

辽宁省节水工作存在问题及对策

辽宁省是水资源严重短缺的省份之一，水资源不足已成为制约全省经济社会可持续发展的瓶颈。通过对全省大型取用水户的调查分析，找出了辽宁省节水工作中存在的问题，提出了节约用水工作的对策和建议。