Unsupervised regionalization of the United States into landscape pattern types

We present a pattern-based regionalization of the conterminous US – a partitioning of the country into a number of mutually exclusive and exhaustive regions that maximizes the intra-region stationarity of land cover patterns and inter-region disparity between those patterns. The result is a discretization of the land surface into a number of landscape pattern types (LPTs) – spatial units each containing a unique quasi-stationary pattern of land cover classes. To achieve this goal, we use a recently developed method which utilizes machine vision techniques. First, the entire National Land Cover Dataset (NLCD) is partitioned into a grid of square-size blocks of cells, called motifels. The size of a motifel defines the spatial scale of a local landscape. The land cover classes of cells within a motifel form a local landscape pattern which is mathematically represented by a histogram of co-occurrence features. Using the Jensen–Shannon divergence as a dissimilarity function between patterns we group the motifels into several LPTs. The grouping procedure consists of two phases. First, the grid of motifels is partitioned spatially using a region-growing segmentation algorithm. Then, the resulting segments of this grid, each represented by its medoid, are clustered using a hierarchical algorithm with Ward’s linkage. The broad-extent maps of progressively more generalized LPTs resulting from this procedure are shown and discussed. Our delineated LPTs agree well with the perceptual patterns seen in the NLCD map.     

High-reliability zircon separation for hunting the oldest material on Earth: An automatic zircon separator with image-processing/microtweezers-manipulating system and double-step dating

Despite the recent development in radiometric dating of numerous zircons by LA-ICPMS, mineral separation still remains a major obstacle, particularly in the search for the oldest material on Earth. To improve the efficiency in zircon separation by an order of magnitude, we have designed/developed a new machine-an automatic zircon separator(AZS). This is designed particularly for automatic pick-up of100 μm-sized zircon grains out of a heavy mineral fraction after conventional separation procedures. The AZS operates in three modes:(1) image processing to choose targeted individual zircon grains out of all heavy minerals spread on a tray,(2) automatic capturing of the individual zircon grains with microtweezers, and(3) placing them one-by-one in a coordinated alignment on a receiving tray. The automatic capturing was designed/created for continuous mineral selecting without human presence for many hours. This software also enables the registration of each separated zircon grain for dating, by recording digital photo-image, optical(color) indices, and coordinates on a receiving tray. We developed two new approaches for the dating; i.e.(1) direct dating of zircons selected by LA-ICPMS without conventional resin-mounting/polishing,(2) high speed U-Pb dating, combined with conventional sample preparation procedures using the new equipment with multiple-ion counting detectors(LA-MIC-ICPMS).With the first approach, Pb-Pb ages obtained from the surface of a mineral were crosschecked with the interior of the same grain after resin-mounting/polishing. With the second approach, the amount of time required for dating one zircon grain is ca. 20 s, and a sample throughput of 150 grains per hour can be achieved with sufficient precision(ca. 0.5%).We tested the practical efficiency of the AZS, by analyzing an Archean Jack Hills conglomerate in Western Australia with the known oldest(4.3 Ga) zircon on Earth. Preliminary results are positive; we were able to obtain more than 194 zircons that are over 4.0 Ga out of ca. 3800 checked grains, and 9 grains were over 4300 Ma with the oldest at 4371 ± 7 Ma. This separation system by AZS, combined with the new approaches, guarantees much higher yield in the hunt for old zircons.     

基于岩石图像深度学习的岩性自动识别与分类方法

岩石岩性的识别与分类对于地质分析极为重要,采用机器学习的方法建立识别模型进行自动分类是一条新的途径。基于Inception-v3深度卷积神经网络模型,建立了岩石图像集分析的深度学习迁移模型,运用迁移学习方法实现了岩石岩性的自动识别与分类。采用此方法对所采集的173张花岗岩图像、152张千枚岩图像和246张角砾岩图像进行了学习和识别分类研究,通过训练学习建立岩石图像深度学习迁移模型,并分别采用训练集和测试集中的岩石图像对模型进行了检验分析。对于训练集中的岩石图像,每组岩石分别用3张图像测试,三种岩石的岩性分类均正确,且分类概率值均达到90%以上,显示了模型良好的鲁棒性;对于测试集中的岩石图像,每组岩石分别采用9张图像进行识别分析,三种岩石的岩性分类均正确,并且千枚岩组图像分类概率均高于90%,但是花岗岩组2张图像和角砾岩组的1张图像分类概率值不足70%,概率值较其他岩石图像低,推测其原因是训练集中相同模式的岩石图像较少,导致模型的泛化能力减小。为了提高识别精确度,对准确率较低的岩石图像进行截取,分别取其中的3张图像加入训练集进行再训练,增加与测试图像具有相同模式的训练样本;在新的模型中,对3张图像进行二次检验,测试概率值均达到85%以上,说明在数据足够的状况下模型具有良好的学习能力。与传统的机器学习方法相比,所提出的岩石图像深度学习方法具有以下优点:第一,模型通过搜索图像像素点提取物体特征,不需要手动提取待分类物体特征;第二,对于图像像素大小,成像距离及光照要求低;第三,采用适当的训练集可获得较好的识别分类效果,并具有良好鲁棒性和泛化能力。     

Research Progress and Prospect on Reactivation of Ancient Landslides

The ancient landslide has endured long-term slope evolution which results in its complicated material and special rock-soil properties. The risk of ancient landslide reactivation is substantially increasing due to the increase of intensified human engineering activities and the frequency of extreme weather events. Many ancient landslides have been reactivated all over the world and led to serious fatalities and severe damage to many important engineering facilities such as transportation and hydropower engineering projects. On the basis of the analysis of the research situation about the ancient landslides at home and abroad, the main research advances were summarized including the regional developing laws and recognizing of the ancient landslides, the mechanics properties of ancient landslide body and related sliding zone, reactivation mechanism of ancient landslides, reactivating process and modeling analysis of ancient landslides, early recognization of ancient landslide reactivation, etc. To meet the demands of disaster prevention and reduction, three key scientific issues were put forward to be solved: ①automaticaly establishing the methodology and identification criterions for recognition of ancient landslide; ②revealing the reactivation mechanism of ancient landslide based on a new strength theory; ③establishing the early rapid recognition method and predictive model for ancient landslide reactivation. Solving the above mentioned scientific theory and methodology will facilitate the planning and site selection of major projects as well as the disaster prevention and reduction in ancient landslide developing areas.     

Novel fusion approach on automatic object extraction from spatial data: case study Worldview-2 and TOPO5000

The automatic extraction of information content from remotely sensed data is always challenging. We suggest a novel fusion approach to improve the extraction of this information from mono-satellite images. A Worldview-2 (WV-2) pan-sharpened image and a 1/5000-scaled topographic vector map (TOPO5000) were used as the sample data. Firstly, the buildings and roads were manually extracted from WV-2 to point out the maximum extractable information content. Subsequently, object-based automatic extractions were performed. After achieving two-dimensional results, a normalized digital surface model (nDSM) was generated from the underlying digital aerial photos of TOPO5000, and the automatic extraction was repeated by fusion with the nDSM to include individual object heights as an additional band for classification. The contribution was tested by precision, completeness and overall quality. Novel fusion technique increased the success of automatic extraction by 7% for the number of buildings and by 23% for the length of roads.     

测井地质学前世、今生与未来——写在《测井地质学·第二版》出版之时

测井地质学以地质学和测井学的方法理论为指导,综合运用各种测井信息,来解决基础地质和石油地质的地质问题。经过数十年发展,测井地质学在油气勘探开发各个环节得到广泛应用。非常规油气的兴起使得测井地质学正面临多重挑战和全新探索,亟需建立针对非常规油气的测井地质学综合方法理论体系。本文以《测井地质学·第二版》出版为契机,系统归纳了测井地质学的起源及发展历程、主要研究内容和研究方法流程。然后总结了测井资料与地震、地质信息的匹配性,并分析了不同探测特性测井方法纵向分辨率区间特征。在此基础上评述了测井地质学在井旁构造解析、沉积学特征研究、层序地层划分、地应力方向判别及大小计算、井壁裂缝识别与评价、烃源岩评价以及非常规油气“七性关系”综合评价当中的应用。但由于测井资料的负载能力有限性、测井与地质信息属性不对应性以及测井资料本身的多解性,使得测井地质学在测井—地质转换、非常规油气测井评价及其与人工智能融合方面还存在一定的问题。因此加强基础岩石物理研究,挖掘测井曲线中包含的地质信息,并与人工智能相结合,将拓展测井地质学研究的精度和广度,从而使其未来可更好地应用至非常规油气测井评价等实践工作中。     

DYYZ-Ⅱ型自动气象站测温故障排除方法与思路

文章通过DYYZ-Ⅱ型自动气象站出现的一次气温、地温数据缺测故障,分析了DYYZ-Ⅱ型自动气象站出现测温故障时的故障排除过程与排除思路,为保障技术人员在处理此类故障时提供了方法和依据。     

基于脑电信号评价情绪状态及听觉激励对改善情绪的研究

如何从脑活动数据中发现具有生物学意义的知识及规律,正成为当今神经信息学理论与实践研究的热点与难点.基于脑电的情绪识别和情绪及其脑机制的关联研究已经成为神经工程和生物医学工程领域的热门课题.本研究基于脑电信号评价情绪状态,尝试借助脑电耳机来监测"毫无头绪"的脑电波,重点开展了听觉激励对改善焦虑情绪状态的实验研究.借助特定中国古风纯音乐产生听觉信号激励,激发出被试者平静放松的情绪状态,从而改善焦虑,发现音乐和情绪之间可能存在的关系.将来在BCI技术研究的基础之上,情绪识别可以进一步满足患有特定疾病（ALS症、脑瘫、脑干中风、脊髓损伤等）的患者人群的某些更高级的需求,即通过脑电分析出其对事物的喜好度,比如音乐的选择,从而设计出更加友善和智能的BCI系统,在医疗康复领域具有很好的应用前景.     

基于镜像图的LRC和CRC偏差结合的人脸识别

为了提高人脸识别率及更好地显示人脸特征,本文提出了一种基于镜像图的LRC和CRC偏差结合的人脸识别方法.该方法首先生成一种镜像人脸,再通过融合原始人脸和镜像人脸形成新的混合训练样本,最后利用LRC和CRC偏差结合进行人脸识别.新方法增加了训练样本的数目,克服了由于光照和姿态等外部因素带来的影响.实验结果表明,镜像图与LRC和CRC偏差结合的人脸识别方法提高了人脸识别的准确性.