基于YOLO的街景影像中行人车辆检测方法

车辆等移动平台搭载全景相机、全球导航卫星系统接收机、惯性测量单元等传感器,可以采集360°的街景.在街景影像中检测行人车辆,一方面可以在数据发布时更好地保护隐私,另一方面可以为城市治理能力提升提供数据支撑.由于街景影像具有表达环境复杂、数据量庞大、目标距离变化大等特点,现有的行人车辆检测算法应用时均存在一定局限.为此,提出一种兼顾效率和精度的街景影像中行人车辆检测方法,利用球形投影原理和先验知识从街景的经纬映射图中划分出目标区域,接着使用你只观察一次(You Only Look Once,YOLO)v4模型从目标区域检测行人和车辆.实验证明,本文方法车辆检测的准确率高于91％,行人检测的准确率高于73％,每张平均耗时21 ms,具备准确率高、速度快的优势,满足实际项目的数据生产需求.     

地球系统动力学模式和模拟研究的进展

概述地球系统动力学模式的由来及国内外目前的研究和模拟进展。地球系统动力学模式是描述全球气候以及生态和环境系统的整体耦合演变的数学表达。利用他作大规模数值模拟,以便认识和预测全球和区域的气候和生态环境变化,有效应对防灾减灾和规划可持续发展等。目前国内外尚未完全研制出可供实用的地球系统模式,还须二三年时间。     

长江口和浙江近岸海域表层沉积物中颗粒磷的形态分布和影响因素

研究长江口和浙江近岸海域的水团流向、缺氧程度、潮汐作用以及养殖活动对表层沉积物中颗粒磷赋存形式的影响,对东海生态环境的可持续发展有着重要的意义。长江口以北和浙江近岸泥质区各存在一个缺氧中心,杭州湾属于强潮区,而象山港遍布养殖场。2018年8—9月在上述4个海域分别采集表层沉积物,采用SEDEX方法进行了颗粒磷的形态划分和含量测定。研究结果表明:在长江口外北部缺氧区和浙江近岸泥质区的南部缺氧区,由于沉积物-水界面的还原环境可以活化铁结合态磷,使之转变为生物可以直接利用的弱吸附态无机磷,因此表现出沉积物中铁结合态磷的低值和弱吸附态无机磷的高值;前者的缺氧程度高于后者,因此其铁结合态磷的含量相应更低。杭州湾和象山港海域均受潮汐作用影响,两者沉积物中磷的分布比较均匀。象山港表层沉积物中铁结合态磷含量均较低,这可能是因为养殖活动造成的水体缺氧通过潮汐活动扩散到整个港底。象山港废弃养殖场沉积物中的碎屑磷和残余有机磷含量显著低于非养殖区,可能是由于养殖活动累积的生物沉积稀释了陆源输入的碎屑颗粒。     

长江口年代际冲淤演变趋势预测与应对策略

受流域来水来沙条件变化及河口大型工程建设的综合影响,长江口呈现新的冲淤格局,为预测未来演变趋势,本研究基于前期研究中建立的长江口年代际冲淤演变预测模型（Delft3D）,未来情景考虑不同来沙量条件和相对海平面上升速率。预测结果表明,到2035年长江口整体以冲刷为主,口内河段主槽和浅滩边缘冲刷较明显,仅高滩局部淤积;到2050年口内河段保持净冲刷状态,拦门沙地区在现状来沙量条件下略有淤积,但在极端低来沙量条件下转变为净冲刷状态,海平面上升对拦门沙地区冲刷具有一定抵消作用,但不会使冲淤状态产生本质改变。本研究分析认为,长江口局部区域未来冲淤趋势可能对河口综合治理与保护产生不利影响,针对新格局条件下的滩槽河势稳定、重要洲滩保护、重大工程安全评估、冲刷致灾研判以及海堤防护标准再评估等方面提出了对策建议,可为新时期长江口综合治理与可持续发展提供参考。     

我国华南3月份降水异常的可能影响因子分析

利用1951～2005年华南地区3月份的降水资料、NOAA海温资料、Ni?o3.4指数和NCEP再分析资料,分析了华南3月份降水异常与同期环流场、全球海温场的关系,从环流和海温的角度揭示了华南3月份降水异常的可能原因。结果表明,当华南3月份降水偏多(少)时,在对流层中低层,北太平洋海区存在气旋(反气旋)性环流异常,西太平洋及南海海面上存在反气旋(气旋)性环流异常,这样的环流异常有利(不利)于东南暖湿气流与北方东部异常冷空气在华南地区形成水汽辐合,导致降水显著增多(减少)。进一步的分析表明,ENSO和北印度洋及南海附近海温是影响华南3月份降水异常的重要外强迫因子,ENSO对华南3月降水异常的影响是通过影响春季西太平洋副热带高压和低层风场异常实现的,而北印度洋及南海附近海温对华南3月降水异常的影响则是通过垂直环流场异常和低层风场以及西太平洋副热带高压异常来实现的。     

On shallow-water bottom reverberation frequency dependence

Determinations of bottom scattering strength in the decade below 1 kHz under downward refracting conditions have been made using acoustic reverberation and transmission data from the 2001 East China Sea Asian Seas International Acoustic Experiment (ASIAEX). The measurements were performed using explosive sources and receiving hydrophones in ship-suspended vertical-line arrays. The focus of this paper has been the dependence of bottom scattering strength on the frequency and characterization of the uncertainties associated with the extraction of scattering strength from reverberation. The derived bottom scattering strength gradually rises with frequency from 100-300 Hz and then more rapidly above 300 Hz. A potential explanation suggests that the frequency variation results from two scattering mechanisms, rough layer scattering at the low end of the band and sediment near-surface volume scattering at the high end. The spatial extrapolation of these results is explored by comparing them with similarly derived scattering strengths using data obtained under the Navy's Harsh Environments Program at a somewhat separated site (56 km) under environmental conditions similar to those during ASIAEX. In the ASIAEX analysis, it has been found that the largest source of uncertainty in the scattering-strength frequency dependence arises from persistence of finite-amplitude effects associated with the source signal.     

An investigation of sand–dust storm events and land surface characteristics in China using NOAA NDVI data

Observations from 560 weather stations in China show that sand–dust storms occur most frequently in April in north China. The region consists of Sub-dry Mid Temperate, Dry Mid Temperate, Sub-dry South Temperate and Dry South Temperate Zones and much of the land surface is desert or semi-desert: it is relatively dry with minimal rainfall and a high annual mean temperature. In most regions of China, the annual mean frequency of sand–dust events decreased sharply between 1980 and 1997 and then increased from 1997 to 2000. Statistical analyses demonstrate that the frequency of sand–dust storms correlates highly with wind speed, which in turn is strongly related to land surface features; on the other hand, a significant correlation between storm events and other atmospheric quantities such as precipitation and temperature was not observed. Accordingly, land surface cover characteristics (vegetation, snowfall and soil texture) may play a significant role in determining the occurrence of sand–dust storms in China. Analysis of Normalized Difference Vegetation Index derived from National Oceanic and Atmospheric Administration and Empirical Orthogonal Function show that since 1995 surface vegetation cover in large areas of Northern China has significantly deteriorated. Moreover, a high correlation is shown to exist among the annual occurrence of sand–dust storms, surface vegetation cover and snowfall. This suggests that the deterioration of surface vegetation cover may strongly influence the occurrence of sand–dust storms in China. Soils with coarse and medium textures are found to be more associated with sand–dust storms than other soils.