Efficient collection of training data for sub-pixel land cover classification using neural networks

Artificial neural networks (ANNs) are a popular class of techniques for performing soft classifications of satellite images. They have successfully been applied for estimating crop areas through sub-pixel classification of medium to low resolution images. Before a network can be used for classification and estimation, however, it has to be trained. The collection of the reference area fractions needed to train an ANN is often both time-consuming and expensive. This study focuses on strategies for decreasing the efforts needed to collect the necessary reference data, without compromising the accuracy of the resulting area estimates. Two aspects were studied: the spatial sampling scheme (i) and the possibility for reusing trained networks in multiple consecutive seasons (ii). Belgium was chosen as the study area because of the vast amount of reference data available. Time series of monthly NDVI composites for both SPOT-VGT and MODIS were used as the network inputs. The results showed that accurate regional crop area estimation (R² > 80%) is possible using only 1% of the entire area for network training, provided that the training samples used are representative for the land use variability present in the study area. Limiting the training samples to a specific subset of the population, either geographically or thematically, significantly decreased the accuracy of the estimates. The results also indicate that the use of ANNs trained with data from one season to estimate area fractions in another season is not to be recommended. The interannual variability observed in the endmembers’ spectral signatures underlines the importance of using up-to-date training samples. It can thus be concluded that the representativeness of the training samples, both regarding the spatial and the temporal aspects, is an important issue in crop area estimation using ANNs that should not easily be ignored.     

养殖水体环境参数随潮汐变化特征及相关性分析——以湛江特呈岛东近岸海域为例

利用2018年4-5月在湛江特呈岛东边海域养殖区内测定的数据,以及海事服务网提供的潮汐数据,分析了研究区域潮汐、流速和流向的变化规律,研究了溶解氧(Dissolved Oxygen,DO)、化学需氧量CODMn(Chemical Oxygen Demand)、二氧化碳CO2浓度和pH等海水环境参数随潮汐变化的特征以及相关性,建立了分别利用DO和pH值来估算海水CO2浓度的新算法,平均估算精度分别达到90%和95%。分析结果表明:(1)研究区域呈现不规则半日潮,流速随潮汐变化分为正相关和负相关两种类型;流向随着潮汐变化基本是东南与西北向;(2) CODMn的变化完全取决于涨落潮过程,涨潮时CODMn值逐渐下降,而落潮过程CODMn值逐渐上升;(3) CO2浓度呈现出明显的日变化,在午后数值达到最低值,不完全受制于潮汐过程的控制;(4) DO和pH两者的变化趋势高度一致,其变化过程与CO2浓度变化过程正好相反;(5)同一地点的海洋环境参数随潮汐涨落有很大的变异性,因而在进行空间大面海洋环境参数调查时,用某时刻的值代表该点的全天情况会出现较大的误差。     

扬子及周缘地区上奥陶统-下志留统烃源岩发育环境及其控制因素

扬子地区下古生界发育了上奥陶统五峰组和下志留统龙马溪组两套有效烃源岩,查清其发育环境及其控制因素具有重要的科学理论和油气勘探意义.晚奥陶世开始,扬子地区进入碎屑岩陆棚演化阶段;五峰期-龙马溪期,扬子地区主体为局限的深水陆棚环境,总体呈现出浅水陆棚、深水陆棚、次深海共存的古地理格局.沉积体系展布和沉积演化主要受扬子陆块与华夏陆块的汇聚作用控制.扬子地区上奥陶统-下志留统烃源岩的形成与生烃母质生物的高生产力和高埋藏率、冰期-冰后期之交的气温快速转暖、海平面快速上升以及粘土矿物在有机质富集保存过程中的赋存驻留作用等密切相关.     

地震-电成像联合探测试验——以淄博市活断层探测为例

在城市浅部活断层探测中,地震和电成像是2种主要的地球物理勘探手段。它们既可独立开展工作,也可在地质构造复杂地段进行联合探测,以更合理地解释探测区域的地质构造。文中利用电阻率和地震纵波速度之间的简单关系,进行了地震-电成像联合反演的初步探索。结果表明,通过参考电成像图像的电阻率分布,可获得合理的准二维地震波叠加速度模型,最终获得探测区域合理的地震时/深剖面,深化对地质构造的认识和提供合理的解释     

Monitoring Contact State of Laboratory Fault with Coda Transmission Waves

We monitored the amplitude changes of coda transmission waves around 500kHz across the frictional interface of a simulated 1.5-meter-long fault during normal stress holding test. We find that the amplitude of coda transmission waves increases with the logarithm of stationary contact time. Localized increase amounted to a level ranging from 4% to 16% along the fault is observed during the 1-hour experiment. We discuss that the frictional strength at mesoscopic scale, which is related to the amplitude of coda transmission waves, is responsible for the phenomenon. Combining the reported method with other complementary approaches will enhance the understanding of fault mechanism either at laboratory or on-site applications.     

宁芜地区中生代岩浆岩的成因——岩石学与Nd,Sr,Pb同位素证据

宁芜中生代火山盆地产出火山岩与侵入岩，火山岩以玄武粗安岩，粗安岩和粗面岩为主，安山岩和响岩少量，火山岩以高钾富碱为特征，已确定为橄榄安粗岩系。侵入岩以辉长闪长玢岩－一辉长闪长岩为主，以高钠低硅为特征，并有辉长岩和花岗岩产生，据地质学和Ｎｄ，Ｓｒ，Ｐｂ同位素资料，侵入岩与火山岩属同一个岩浆系列，是碱性玄武岩浆在下地壳经过轻度ＡＦＣ混合后，侵入上地壳，在轻度混染的情况下，经过以结晶分离为主的岩浆分异形     

海相黏性土状态与直接快剪试验指标的经验关系

通过对大连海域第四系全新统海相黏性土385组试验数据的统计与分析,建立了第四系全新统海相黏性土(Q4m)液性指数IL与直接快剪试验指标-黏聚力Cq、内摩擦角q的经验关系,即Cq=10.623 I-1.8114L; q=-3.8289ln(IL)+5.8634。     

黑龙江密山地区晚石炭世含煤岩系沉积和构造特征

从建造和改造两个方面,系统研究了晚石炭世珍子山组的发育特征,分析了与赋煤有关的地质因素。结果表明:珍子山组含煤段沉积厚度比较稳定,含煤性较好,受后期构造破坏较轻,是较好的赋煤区。该研究为密山地区进一步找煤指出了方向。     

Areal extent of the 1988 flood in Bangladesh: How much did the satellite imagery show?

This paper presents the methods and results of visual interpretation of satellite imagery for estimating areal extent of the 1988 flood in Bangladesh. The main limitation of this method was that cloud-free images were not available for the peak flood period in the first week of September. Relatively cloud-free images on three selected dates in the last three weeks of September showed that areas of inundation on those dates ranged from 44000 to 60000 km² ± 5% (31 to 42% of Bangladesh). The Government of Bangladesh official estimates of the areal extent of flooding were tentative in nature, since these were based on a choropleth map of relative proportions (percentages) of flooding in different parts of the country. The official estimate of the maximum extent of flooding was 82000 km² (57% of Bangladesh), i.e. 15 to 26% larger than the area shown on satellite imagery. The actual extent of flooding was certainly larger than the estimates from satellite imagery but, perhaps, smaller than the official estimates, since a reduction of 15 to 26% of flood area in one to three weeks seemed unlikely. An alternative method of mapping flood-affected areas by using newspaper-interpreted data was attempted, but the method had limited values because of reporting bias.     

山东济宁组群结构大城市地下水资源评价和应急供水探讨

阐述了济宁组群结构大城市地下水资源量与开发利用现状,指出因开采地下水产生的环境地质问题,根据现有水源地的开采潜力分析,提出河里、双庙、石桥、大南铺-新驿和济宁城区水源地适宜作为应急供水水源,并采用不同方法计算了应急供水总量.当济宁市城区水源地无法供水时,在无新增供水水源的情况下,其它应急水源地的供水保证程度(按2010年规划需水量计算)为60%,可基本满足城市生活供水需求.