用人工神经网络实现地震记录中的废道自动切除

切除废道（即道编辑）的工作是地震信号处理中重要而又工作量巨大的一步.本文提出了一种用人工神经网络实现的自动切除方法,所用的神经网络模型为新奇滤波器模型,方法简单实用,切除效果令人满意.文中介绍了方法的原理并给出了几个实验结果.     

东北平原农田防护林网区土地资源系统结构及动态

本文在对东北平原农田防护林网区土地资源系统基本结构分析的基础上,利用典型区域的数据对该土地资源系统动态进行分析。本区的土地资源系统进入广泛的农业开发时间仅五、六十年,但经过强人工作用,已从垦殖期转入衰竭期。只有以人工再生(高投入)的方式才能保证土地资源的永续利用,维持系统的良性循环。否则,仅仅依赖自然再生(低投入)为主要的物质能量返回形式,必然导致粮畜生产带走物质能量,造成土地肥力得不到必要补充,形成土地资源日益衰竭的恶性循环系统。     

Centuries of artificial recharge on the southern edge of the Sierra Nevada (Granada,Spain)

In the Alpujarra (southern Spain), every year between the March and June, an ancestral practice continues in the form of diverting water from the rivers by way of an extensive network of irrigation channels (acequias) to well-defined, highly permeable areas. This practice, known ascareos, constitutes an ancient example of artificial recharge. The objective is to guarantee a supply of drinking water during the dry months, as well as improve the physicochemical characteristics of the water. In addition, this system helps maintain moisture in the immediate environment, and thus has a positive effect on local vegetation.     

Quantitative analysis of springs

Growing demand for groundwater resources and stringent environmental concerns has led to large groundwater investigations, including characterization of aquifer systems that are hydraulically connected to springs. A pumping test is one of the most reliable means of quantifying hydraulic characteristics and the response of natural springs to pumping because it yields results that, in general, are representative of a larger area than are results from a single point observation. Recharge to the aquifer sustaining discharge from springs must be evaluated prior to the utilization of springs. The spring hydrograph is analyzed, as the shape of a hydrograph is a reflection of the response of the aquifer to recharge. The form and rate of recession provide significant information on the storage, lithological composition, and structural characteristics of the aquifer system sustaining the spring. Water tracing techniques have been developed and used over a period of centuries to delineate catchment boundaries, estimate groundwater flow velocities, determine areas of recharge, and identify sources of pollution of spring water.     

黄土孔隙性分类判别的人工神经网络方法

运用人工神经网络的一典型模型——“反向传播”神经网络,对洛川黄土孔隙性的实测数据进行了分析,建立了洛川黄土孔隙性预测的计算机智能专家系统。结果表明,神经网络方法性能良好,可望成为黄土孔隙性分类、判别的有效辅助手段。     

祁连山空中云水资源开发潜力研究新进展

祁连山及河西走廊地区是我国气候变化敏感区和生态脆弱区。本文通过对2005a以来祁连山云水资源相关最新研究成果的总结,旨在揭示气候变暖对祁连山云水资源影响包括气候变化、水汽时空变化、地形云特征的基础上,分析评估了祁连山云水资源开发潜力及效益。实验研究表明,在祁连山区开展人工增雨(雪)作业,每年可增加降水量3.7×10_⁸～７.4×10_⁸m_³。近10年来,甘肃春季飞机人工增雨作业年均增加降水量为13.5×10_⁸m_³,平均投入产出比为1:30。     

An Artificial Intelligent design for GPS Surveying Networks

Metaheuristic techniques, which are based on ideas of Artificial Intelligence, are among the best methods for solving computationally the GPS surveying network problem. In this paper, the ant colony optimization metaheuristic, which is inspired by the behavior of real ant colonies, is developed to efficiently provide a general framework for optimizing GPS surveying networks. In this framework, a set of ants co-operate together using an indirect communication procedure to find good GPS observation schedules. A GPS surveying network can be defined as a set of stations, co-ordinated by a series of sessions formed by placing receivers on the stations. The problem is to search for the best order in which to observe these sessions to give the best schedule at minimum cost. Computational results obtained by applying the proposed technique on several networks, with known and unknown optimal schedules, prove the effectiveness of the proposed metaheuristic technique to solve the GPS surveying network problem.     

Biomonitoring of coastal pollution status using protozoan communities with a modified PFU method

Structural and functional parameters of protozoan communities were assessed as indicators of water quality in Korean coastal waters in the summer of 2000. A modified polyurethane foam unit (PFU) method, named the bottled PFU (BPFU) system, was used in order to carry out the bioassessment. Both parameters suggested that biomonitoring using the BPFU system was more effective than the conventional PFU method in offshore areas. The species number collected by the BPFU system generally decreased as pollution intensity increased at three main stations and was greater than that collected using the PFU method (paired t-test, t = 4.83, p < 0.0001). The Margalef diversity index coincided well with the water conditions. The diversity index values calculated from the BPFU system were also significantly higher than those from the PFU method (paired t-test, t = 5.37, p < 0.0001). Furthermore, the functional parameters, i.e. S(eq),G and T90%, correlated with the pollution status and could thus clearly discriminate the different classes of water quality.     

区域降水数值预报产品人工神经网络释用预报研究

利用T213、日本细网格降水预报等数值预报产品,采用人工神经网络方法进行预报释用。通过聚类分析方法对广西自治区测站进行分类,简化预报对象,对数量众多的T213数值预报产品采用自然正交分解(EOF)方法,浓缩大量因子的有效信息,并结合日本降水预报因子建立广西5~6月区域降水量级的逐日人工神经网络预报模型。运用与实际业务预报相同的方法进行逐日预报试验。结果表明,用这种数值预报产品释用方法建立广西3个预报区域的B-P人工神经网络预报模型对中雨以上降水量级预报的TS评分分别为0.55、0.5和0.26,比目前业务预报中参考使用的T213和日本数值预报产品降水预报具有更好的预报效果。     

一种基于人工免疫的图像分割算法

图像二维熵分割,一直因耗时长而限制了实际应用。本文借鉴生物免疫思想,提出二维熵图像分割的人工免疫算法。在克隆选择算法中引入疫苗的免疫接种,用于优化最优分割阈值对的搜索过程。在遥感高分辨率图像上的实验显示,该算法不仅能准确搜索到最优阈值对,而且计算时间只有传统算法的1.8%。该算法也验证了人工免疫思想用于图像分割的可行性和有效性。