基于热红外遥感数据和光谱混合分解模型的城市不透水面估算

不透水面的迅速增长是城市化的显著特征之一,针对大范围的城市监测,运用遥感技术迅速提取城市不透水面是当前国内外研究的热点。论文选用Landsat 7的ETM+影像,基于光谱混合分解模型,结合热红外遥感数据反演生成的地表温度,研究杭州市的不透水面分布信息的提取。通过高、低反照率、植被及土壤4类光谱端元的线性组合来表征不同城市土地类型,并利用地表温度和土壤分量分别剔除高、低反照率分量中的"噪声",综合修正后的高反照率分量和低反照率分量估算杭州市不透水面分布。结果显示,研究区中均方根误差的平均值为0.003 6,不透水面分布结果与同期Google earth上的高分辨率影像和SPOT 4影像的解译结果对比分析,绝大多数样本的估算值与解译值之差落在±0.15区间内,精度令人满意。研究表明,热红外遥感数据和光谱混合分解模型相结合,可以实现对不透水面进行快速、精确的估算。     

Preliminary studies of video images of smoke dispersion in the near wake of a model building

A summary of analyses of video images of smoke in a wind tunnel study of dispersion in the near wake of a model building is presented. The analyses provide information on both the instantaneous and time-average patterns of dispersion. Since the images represent vertically-integrated or crosswind-integrated smoke concentration, only the primary spatial and temporal scales of pollutant dispersion can be examined. Special graphic displays of the results are presented to assist in the data interpretation. The video image format is shown to have great potential as an easily quantifiable electronic medium for studying the dispersion of smoke.     

基于深度学习的近岸海浪等级视频监测

深度学习是机器学习的重要研究领域,同时作为大数据的有效处理和分析工具越来越受到关注。以多源长时间序列近岸海浪视频环境数据为样本,波浪仪同步测量海浪等级数据为图像标签,构建了面向海洋环境适用于深度学习的海浪训练集、测试集。通过数据扩增技术对视频监测数据进行预处理,提高模型泛化能力,依据视频的相关性,引入误差函数,优化模型灵敏度,提出了适用于海洋领域海浪等级深度学习模型架构（Wave-CNNs）,最后将提出的改进深度学习模型应用于3000样本海浪图像训练集,并通过300样本海浪图像测试集对结果进行验证,实验结果表明,算法对3个等级海浪识别精度达到了66.6%,优于传统Bayes及SVM方法。     

基于非负矩阵分解和支持向量机的墨西哥湾溢油HJ-1星遥感图像分类

为提高国产自主卫星在海洋溢油监测中的精度,提出一种基于非负矩阵分解和支持向量机的环境减灾星(HJ-1)海洋溢油遥感图像分类算法。针对HJ-1星遥感图像,首先利用非负矩阵分解算法进行图像特征提取,相对图像光谱和纹理等图像基本特征,构造具有针对性的溢油图像局部化非负特征,更符合遥感图像特征所对应的物理意义。进而在新特征的基础上,采用支持向量机实现遥感图像分类,解决小样本训练问题。通过墨西哥湾溢油遥感图像仿真实验比较,证明该方法在HJ-1星溢油图像分类中的有效性。     

Two-dimensional spectrophotometry and pseudo-color representation of chemical reaction patterns

The formation of spatio-temporal patterns in layers of chemical and biochemical media is observed and quantified by a two-dimensional spectrophotometer based on a computerized video equipment with high spatial, temporal, and intensity resolution. Images are shown, to some extent, in pseudo-colors. The use of pseudo-colors and three-dimensional perspective representations enhances the perception of the information contained in an image and the inherent order of the spatial structures.     

水质综合评价的投影寻踪模型

为检验水质评价标准的合理性，解决各单项水质指标评价结果的不相容问题，提高水质评价模型的分辨率，提出了一种新的模型－投影寻踪模型，它的评价结果是实数值。实例的计算结果表明，该方案是有效的和通用的，可应用于各种环境质量综合评价中。     

A CONSERVATION ASSESSMENT OF CENTRAL AMERICAN VEGETATION AND ECOREGIONS

Acomprehensive,standardized,andthematicallyappropriatemapofCentralAmericanvegetationandlandcovertypeswasdevelopedbyclassifyingtheremotesensedimagery(AVHRR)1km²     

Effects of spatial resolution of remotely sensed data on estimating urban impervious surfaces

Impervious surfaces are the result of urbanization that can be explicitly quantified, managed and controlled at each stage of land development. It is a very useful environmental indicator that can be used to measure the impacts of urbanization on surface runoff, water quality, air quality, biodiversity and microclimate. Therefore, accurate estimation of impervious surfaces is critical for urban environmental monitoring, land management, decision-making and urban planning. Many approaches have been developed to estimate surface imperviousness, using remotely sensed data with various spatial resolutions. However, few studies, have investigated the effects of spatial resolution on estimating surface imperviousness. We compare medium-resolution Landsat data with high-resolution SPOT images to quantify the imperviousness in Beijing, China. The results indicated that the overall 91% accuracy of estimates of imperviousness based on TM data was considerably higher than the 81% accuracy of the SPOT data. The higher resolution SPOT data did not always predict the imperviousness of the land better than the TM data. At the whole city level, the TM data better predicts the percentage cover of impervious surfaces. At the sub-city level, however, the ring belts from the central core to the urban-rural peripheral, the SPOT data may better predict the imperviousness. These results highlighted the need to combine multiple resolution data to quantify the percentage of imperviousness, as higher resolution data do not necessarily lead to more accurate estimates. The methodology and results in this study can be utilized to identify the most suitable remote sensing data to quickly and efficiently extract the pattern of the impervious land, which could provide the base for further study on many related urban environmental problems.     

Prediction of Temperature in Vacuum Arc Remelting in the Presence of Strong Disturbances

A system for predictive, noninvasive temperature measurement in a broad class of manufacturing processes is presented. The system employs a comprehensive, three-tier signal processing algorithm to achieve high accuracy and reliability in the presence of strong disturbances. Temperature estimates are generated by ‘unstructured, shallow knowledge’ based algorithms. The estimation process employs a set of robust features obtained by signal processing that involves constitutive models representing ‘deep knowledge’ about the process dealt with. The application of these models is combined with ‘structured, shallow knowledge’ based techniques. Results obtained in application to the investment casting of titanium, which employs the vacuum arc remelting (VAR) process, are presented.     

An integrated CNC accumulation system for automatic building-around-inserts

In this paper a non-layer-based additive manufacturing (AM) process named computer numerically controlled (CNC) accumulation process is presented for applications such as plastic part repairing and modification. To facilitate the CNC accumulation process, a novel three-dimensional (3D) laser scanning system based on a micro-electro-mechanical system (MEMS) device is developed for in situ scanning of inserted components. The integration of the scanning system in the CNC accumulation process enables the building-around-inserts with little human efforts. A point processing method based on the algebraic point set surface (APSS) fitting and layered depth-normal image (LDNI) representation is developed for converting the scanning points into triangular meshes. The newly developed 3D scanning system is compact and has sufficient accuracy for the CNC accumulation process. Based on the constructed surface model, data processing operations including multi-axis tool path planning and motion control are also investigated. Multiple test cases are performed to illustrate the capability of the integrated CNC accumulation process on addressing the requirements of building-around-inserts.     

Smooth surface fabrication in mask projection based stereolithography

The surface finish is critical for applications such as optics, micro-fluid flow and mechanical assembly, in which optical lenses, fluidic channels and rotating components are all required to be smooth. However, the stair-stepping effect is well known in the layer-based additive manufacturing processes, in which a three-dimensional model is approximated by a set of two-dimensional layers. Consequently the fabricated surfaces have poor surface finishes especially those that are close to the horizontal plane. In this paper, a novel approach for achieving improved surface finish is presented for the mask-image-projection-based stereolithography (MIP-SL) process. Theoretical models and parameter characterization are presented for the developed methods. Accordingly, the process planning and calibration approaches for fabricating smooth up-facing surfaces in the MIP-SL process have been developed. Multiple test cases based on various types of curved surfaces have been performed. A comparison of the build results based on the traditional and the newly developed approaches illustrates the effectiveness of our method.     

智能巡检机器人在自然环境试验中的应用探讨

目的 探索建立自然环境试验智能巡检系统,使用机器人替代部分人工巡检工作。方法 通过分析巡检作业环境和巡检任务,确定智能巡检机器人的功能要求和技术参数,进而开展机器人本体结构设计,融合控制后台、远程数据库等设备,实现智能巡检系统构建。结果 开展了智能巡检数据采集的使用验证,巡检机器人采集的样品图像对焦清晰、分辨率高,样品在图像中的位置大小合理,腐蚀特征展示清楚。巡检系统可对样品图像进行自动分割和智能识别,对蚀点、流痕等损伤特征识别准确,精度可达0.1%。结论 智能巡检系统可以满足自然环境试验的巡检要求,替代人工完成部分巡检工作,智能巡检数据便于入库和使用挖掘,对于自然环境试验的数字化转型有积极的推动作用。     

MBR工艺在污水处理厂提标扩建工程中的应用与优化建议

丹江口某污水处理厂提标扩建工程设计规模为3.0×104 m3/d,出水水质主要指标执行GB 3838—2002《地表水环境质量标准》Ⅳ类,其中ρ(TN)≤15 mg/L。二级处理单元采用A2/O+MBR工艺,处理效果良好,达到设计目标。结合本工程MBR膜池及设备间的设计,介绍了BIM(建筑信息模型)技术在MBR工艺设计中的应用,并根据分析结果对设计中潜在的问题给出优化建议。同时,针对现场实际运行中暴露的问题提出解决方法,如设备布置与起吊、污泥沉降与曝气、回流泵的汽蚀防护等,为其他类似工程提供参考。     

污染土地置换开发前期地形图测绘技术研究

传统测绘技术进行污染土地置换开发前期地形图测绘存在精准度低,整合检查效果差的问题,提出基于空三加密技术的污染土地置换开发前期地形图测绘技术。根据地形勘测结果和土地污染现状,收集基本控制测量数据,以测量数据为基础设计技术实施流程;采用无人机进行航空摄影,获取地形航带分布情况,测量区域范围;利用空三加密测绘技术对数字测图数据进行采集、等高线提取与编辑、地形立体测绘和图像编辑,完成地形图的测绘。     

长江苏州段岸线资源利用遥感调查与GIS分析评价

了解岸线资源利用现状,科学、合理地评价岸线资源条件是实施沿江开发战略、建立沿江基础产业带、科学合理地确定沿江产业布局的前提和基础。利用高分辨率卫星影像,通过影像融合,快速发现目标靶区,精确确定岸线占用单位的空间位置与几何形状;运用GIS分析技术,弄清长江岸线江苏段的资源与利用现状,选取科学合理的评价单元,对研究区内的岸线资源进行评价,为长江岸线苏州段的合理利用与开发提供重要的参考依据。     

钻井过程人员异常行为视频智能识别系统

针对钻井现场视频监控人工值守工作量大、异常行为识别不及时、海量视频数据缺乏实时报警等问题,结合钻井作业操作规程建立井控坐岗监测、起下钻、甩钻具等钻井作业场景典型异常行为判别规则,在基于卷积神经网络的目标检测方法基础上,引入SIFT方法和基于强化学习的困难样本筛选策略,提高了算法识别的准确率与鲁棒性。测试结果表明,在像素分辨率1 920×1 080的情况下,该方法整体识别准确率达到85%以上,平均识别速度为135 ms/帧。在此基础上,开发了钻井过程人员异常行为智能识别系统,试点应用结果表明,该系统能够对接入的视频流进行实时检测,及时、准确识别出坐岗人员擅自离岗、起下钻及甩钻具过程人员危险站位等异常行为,有效提升钻井现场安全管理水平。     

Dystrophin protein and rflp analyis for fetal diagnosis and carrier confirmation of duchenne muscular dystrophy

A pregnant woman with indeterminate Duchenne muscular dystrophy (DMD) carrier status, but with DMD diagnosed in her deceased brother (unavailable for study), presented for prenatal diagnosis, intending to continue the pregnancy only if proven unaffected with DMD with near absolute certainty. Creatine kinase (CK) assays to clarify carrier status were inconclusive. Male sex in the fetus was identified, but DNA restriction fragment length polymorphism (RFLP) analysis was not yet available to this centre to investigate the possible transmission of the DMD gene, and the pregnancy was terminated. Tissue histology and dystrophin protein analysis demonstrated the absence of DMD. In a situation with proven maternal carrier status, future fetal inheritance of the opposite maternal X chromosome would indicate the presence of DMD. However, maternal carrier status remained in doubt through a second pregnancy, even with RFLP studies, and was finally established when dystrophin analysis confirmed the presence of DMD in the second fetus. Histologic findings are presented, contrasting features in the two fetuses. The value of dystrophin analysis for establishing the diagnosis of fetal DMD, in this case proving maternal carrier status in a difficult situation, and for demonstrating DMD gene:RFLP haplotype relationships is illustrated.     

铝合金的实验室盐雾试验腐蚀行为图像特征提取

目的提取铝合金材料在盐雾箱试验中可以表征腐蚀程度的表面形貌图像特征量。方法首先,对采集于表面化学清洗过的试样原始图像进行图像增强等预处理,突出腐蚀部位;其次,基于数字图像处理的分形和小波分解方法提取出分形维数和小波能量特征值。结果与以质量损失量为基础的腐蚀深度特征值进行相关性对比,基于图像分析的特征提取法的准确度和精度比较高。结论该方法可以应用于对试样的腐蚀程度进行定性和定量分析,判断并预测试样的腐蚀速度。     

基于改进光谱角法的红树林高分遥感分类方法研究

传统影像分类方法多利用影像端元光谱进行地物分类,影像的空间结构信息被忽视,本研究结合面向对象分类方法思想以提高红树林遥感分类精度。本研究提出了一种结合端元类型选择、像元提纯等混合像元分解手段及分水岭图像分割算法的改进光谱角影像分类方法,并以山口红树林国家级自然保护区为研究区,利用GF-1号遥感影像数据,在光谱特征分析和地面调查的基础上,对红树林生态系统进行分类,并对分类精度进行分析。研究结果表明:改进的光谱角分类方法对GF-1影像分类效果较好,既兼顾地类光谱组成较复杂时的特殊性,又有效避免结果的破碎化现象,且总体精度达到95%（KAPPA系数0.944）,证明了其在红树林遥感影像分类及信息提取方面的应用潜力,为红树林生态系统业务化遥感监测奠定了基础。     

基于改进电法装置和深度神经网络的填埋场渗滤液水位探测

渗滤液水位检测和管控是填埋场环境风险管控的关键之一,高密度电法装置具有无损、快速和分辨率高等优点,但填埋场铺设的HDPE防渗膜的高阻特性掩盖了渗滤液的低阻特性,反演结果存在探测精度差、探测范围有限等问题,无法有效识别渗滤液水位. 为此,本文设计了改进的川字型采集装置并配套提出了基于深度神经网络的非线性反演算法——EConvNet-C,通过构建高仿真模拟模型获得了代表性场景下的学习样本,并进行学习训练得到了渗滤液水位与观测数据的非线性映射关系. 对上述改进的电法装置和配套的反演算法的有效性进行了验证,并开展了实际填埋场地的案例研究. 结果表明:①基于川字型装置的EConvNet-C反演的电阻率存在明显的分层特征,在场景A中EConvNet-C反演算法的均方误差(MSE)均在0.00230以下,渗滤液水位探测精度均在90.0%以上. ②以“浅层滞水”形式非正常积存情况下,EConvNet-C探测精度略有下降,但MSE仍在0.00420以下,水位探测精度仍在80.0%以上. ③基于传统探测装置的深度神经网络耦合方法得到的渗滤液水位探测精度略低于基于川字型装置的EConvNet-C,而基于传统探测装置的线性反演耦合方法在HDPE膜高阻特性的影响下无法有效识别渗滤液水位. 因此,基于川字型装置的EConvNet-C在填埋场渗滤液水位探测领域具有较大潜力,可为后期开展垃圾填埋场的性能和风险评估提供指导.