共查询到20条相似文献，搜索用时 46 毫秒
1.
本文基于Langston的震源机制分解方法以及Minson和Dreger修正后的近场位移的解析解表达式,给出了一种区域纯偏量矩张量的求解方法,方法使用基于Hilbert变换的波形包络作为增强约束．对该方法进行了数值试验. 试验结果表明,定位准确且速度结构已知的情况下,背景高斯噪声水平达30%时,该方法仍可以得到较稳定的解;使用0.01——0.1 Hz的中长周期频段进行反演,在随机震相到时误差最大plusmn;3 s的情况下,近90%的情况可以得到相对准确的解. 速度结构模型的干扰试验表明,平均模型接近真实地壳结构时,可以得到接近真实的解;模型与真实的地壳结构的误差在10%以内时,得到的解与真实解比较接近;误差接近20%时,矩张量解会出现很大的偏差;地震的定位误差主要影响震源深度的判定,从而导致最终解中出现虚假的非双力偶成分,但在约束为偏量矩张量的情况下,反演得到的断层面参数与真实解偏差不大;在台站方位角覆盖较差(＜60deg;)的情况下,使用平均模型仍可以得到较为准确的解．使用该方法研究了2010年10月24日河南周口太康MW4.7地震的震源机制,结果显示矩张量解与初动符号解较为一致． 相似文献
2.
Rupture model of the 2013 M_W 6.6 Lushan (China) earthquake constrained by a new GPS data set and its effects on potential seismic hazard 
下载免费PDF全文
下载免费PDF全文 Vertical records are critically important when determining the rupture model of an earthquake, especially a thrust earthquake. Due to the relatively low fitness level of near-field vertical displacements, the precision of previous rupture models is relatively low, and the seismic hazard evaluated thereafter should be further updated. In this study, we applied three-component displacement records from GPS stations in and around the source region of the 2013 MW6.6 Lushan earthquake to re-investigate the rupture model.To improve the resolution of the rupture model, records from both continuous and campaign GPS stations were gathered, and secular deformations of the GPS movements were removed from the records of the campaign stations to ensure their reliability. The rupture model was derived by the steepest descent method(SDM), which is based on a layered velocity structure. The peak slip value was about 0.75 m, with a seismic moment release of 9.89 × 10~(18) N·m, which was equivalent to an M_W6.6 event. The inferred fault geometry coincided well with the aftershock distribution of the Lushan earthquake. Unlike previous rupture models, a secondary slip asperity existed at a shallow depth and even touched the ground surface. Based on the distribution of the co-seismic ruptures of the Lushan and Wenchuan earthquakes, post-seismic relaxation of the Wenchuan earthquake, and tectonic loading process, we proposed that the seismic hazard is quite high and still needs special attention in the seismic gap between the two earthquakes. 相似文献
3.
Existing empirical models for estimating liquefaction-induced lateral spread displacement (DLL) have been derived from a dataset poorly distributed with respect to earthquake magnitude and source distance, and also produced from different tectonic source types and faulting mechanisms. Both the poor distribution and mixed tectonic source types and faulting mechanisms of the data have an adverse impact on the reliability of the empirical models. To overcome these problems in the development of empirical models, we replace the direct use of magnitude and source distance with pseudo-displacement derived from spectral acceleration attenuation models that are well supported by earthquake data, and use a modification factor to account for effects of the non-linear soil response. Attenuation models derived from very large and reasonably well-balanced datasets have been selected, one being a Japanese attenuation model and other being a combination of the Sadigh et al. model and the Youngs et al. model. These models are capable of accounting for the effects of earthquake tectonic source type and faulting mechanism. We determined the model coefficients by selecting the pseudo-displacements calculated for a number of spectral periods to achieve an unbiased distribution of residuals with respect to earthquake magnitude and source distance. Sensitivity analyses have been carried out based on the new and existing models, and show that the new model is more robust than the existing models. Comparison with a limited number of data from the 1997 Kocaeli, Turkey earthquake suggests our model provides comparable liquefaction-induced lateral displacement DLL estimates. 相似文献
4.
Fulu Men 《地震工程与工程振动(英文版)》2002,1(2):281-291
In this paper, mantle circulation flow, continental drift, earthquake origin and other mechanical principles are examined
as they apply to earthquake engineering, seismology and dynamics of fluid saturated porous medium. The relationship of mantle
flow to earthquakes is examined and clarified, and a new model, different from Haskell’s, is proposed for the earthquake mechanism.
The proposed new model is based on the discovery that two pairs of jump stress and jump velocity will start to act from the
fault plane. Records obtained directly from recent earthquakes nearby and right on the fault break show a very large velocity
impulse, which verify, indirectly, the new mechanism proposed by the author. Further, at least two physical parameters that
characterize the seismic intensity must be specified, because according to the discontinuous (jump) wave theory, at the earthquake
source, the stress jump and the velocity jump of particle motion should act simultaneously when a sudden break occurs. The
third key parameter is shown to be the break (fracture) propagation speed together with the break plane area. This parameter
influences the form of the unloading time function at the source. The maximum seismic stress in and displacement of a building
are estimated for two unfavorable combinations of the building and its base ground in terms of their relative rigidity. Finally,
it is shown that Biot’s theory of wave propagation in fluid saturated porous media is valid only when fluid flow cannot occur. 相似文献
5.
Deviation of earthquakes from the double‐couple mechanism is an important, but delicate tool to study their source processes. For assessing the double‐couple percentage, the paper suggests to complement the standard least‐square moment‐tensor retrieval with a hierarchic spatio‐temporal grid search, progressively closer to the true source position and time. It enables identification of the double‐couple percentage convergence, while its limit is the resulting double‐couple percentage value, or range. The so‐called double‐couple percentage (DC%) versus correlation plots are introduced and difficulties of the double‐couple percentage assessment are discussed. It is proved that even close to the true source position, where the strike‐dip‐rake angles are already stable (within a few degrees), the double‐couple percentage may still vary by dozens of per cent. Moreover, even at the optimum spatial position, the double‐couple percentage estimate is extremely sensitive (0 to 100%) to small variations of the subevent origin time. This behaviour is explained in terms of the source complexity, implying a time‐dependent moment tensor. Therefore, the double‐couple percentage of complex events depends on the studied frequency band and, in general, also on the station azimuth. This explains broad variations of the double‐couple percentage reports among seismic agencies. Three earthquakes of mutually close epicentres were investigated (Zakynthos, Western Greece, April 2006, magnitudes ～5.5) and a strong non‐double‐couple component of one of them was identified (double‐couple percentage of about 20%). Two equivalent models of this earthquake were found: a single‐event non‐ double‐couple model, and a double‐event model consisting of two double‐couple sources with highly different mechanisms. 相似文献
6.
J. R. Murphy I. O. Kitov B. W. Barker D. D. Sultanov 《Pure and Applied Geophysics》2001,158(11):2077-2101
—?During the period 1965 to 1988, the former Soviet Union (FSU) conducted over 120 peaceful nuclear explosions (PNE) at locations widely dispersed throughout the territories of the FSU. These explosions sample a much wider range of source conditions than do the historical explosions at the known nuclear test sites and, therefore, seismic data recorded from these PNE tests provide a unique resource for use in deriving improved quantitative bounds on the ranges of seismic signal characteristics which may require consideration in global monitoring of the Comprehensive Test-Ban Treaty (CTBT). In this paper we summarize the results of a detailed statistical analysis of broadband seismic data recorded at the Borovoye Geophysical Observatory from 21 of these PNE tests at regional distances extending from about 7 to 19 degrees, as well as the results of theoretical waveform simulation analyses of near-regional (Δ?相似文献
7.
Mine Betül Demircioğlu Karin Şeşetyan Tamer Y. Duman Tolga Çan Senem Tekin Semih Ergintav 《Bulletin of Earthquake Engineering》2018,16(8):3399-3438
Over the years, several local and regional seismic hazard studies have been conducted for the estimation of the seismic hazard in Turkey using different statistical processing tools for instrumental and historical earthquake data and modeling the geologic and tectonic characteristics of the region. Recently developed techniques, increased knowledge and improved databases brought the necessity to review the national active fault database and the compiled earthquake catalogue for the development of a national earthquake hazard map. A national earthquake strategy and action plan were conceived and accordingly with the collaboration of the several institutions and expert researchers, the Revision of Turkish Seismic Hazard Map Project (UDAP-Ç-13-06) was initiated, and finalized at the end of 2014. The scope of the project was confined to the revision of current national seismic hazard map, using the state of the art technologies and knowledge of the active fault, earthquake database, and ground motion prediction equations. The following two seismic source zonation models are developed for the probabilistic earthquake hazard analysis: (1) Area source model, (2) Fault and spatial smoothing seismic source model (FSBCK). In this study, we focus on the development and the characterization of the Fault Source model, the background spatially smoothed seismicity model and intrinsic uncertainty on the earthquake occurrence-rates-estimation. Finally, PSHA results obtained from the fault and spatial smoothed seismic source model are presented for 43, 72, 475 and 2475 years return periods (corresponding to 69, 50, 10, and 2% probability of exceedance in 50 years) for PGA and 5% damped spectral accelerations at 0.2 and 1.0 s. 相似文献
8.
利用山东台网记录的长岛震群2017年2月14日—9月1日期间的波形与震相资料研究长岛地区非弹性衰减系数,得到该地区介质平均Q值与频率f的关系式为Q(f)=363.9f1.374 1。采用Moya等[1]提出的利用遗传算法联合反演得到长岛周边台站的场地响应,根据Brune模型震源参数计算公式求解长岛震群序列地震震源参数。结果显示,各个震源参数之间均存在一定的相关关系,地震矩随ML震级的增大而增加,地震矩与破裂半径R之间存在半对数关系,拐角频率fc随地震矩的增大而减少;长岛地震序列的应力降数值普遍偏小,最大不超过0.9MPa,这意味着长岛震源区整体构造应力较低,也可能指示长岛震群为低摩擦应力的断层作用;震源参数随时间的变化方面,整体而言,长岛震群地震应力降变化起伏很大,在M4.1地震发生前,拐角频率与应力降均发生快速下降后随即翻转上升的现象,证明在M4.1地震发生前震源区整体应力的挤压逐渐增强。 相似文献
9.
We present a simplified method to simulate strong ground motion for a realistic representation of a finite earthquake source burried in a layered earth. This method is based on the stochastic simulation method of Boore (Boore, D. M., 1983, Bull. Seism. Soc. Am. 73, 1865–1894) and the Empirical Greens Function (EFG) method of Irikura (Irikura, K., 1986, Proceedings of the 7th Japan Earthquake symposium, pp. 151–156). The rupture responsible for an earthquake is represented by several subfaults. The geometry of subfaults and their number is decided by the similarity relationships. For simulation of ground motion using the stochastic simulation technique we used the shapping window based on the kinetic source model of the rupture plane. The shaping window deepens on the geometry of the earthquake source and the propagation characteristics of the energy released by various subfaults. The division of large fault into small subfaults and the method for accounting their contribution at the surface is identical to the EGF. The shapping window has been modified to take into account the effect of the transmission of energy released form the finite fault at various boundaries of the layered earth model above the source. In the present method we have applied the correction factor to adjust slip time function of small and large earthquakes. The correction factor is used to simulate strong motion records having basic spectral shape of 2 source model in broad frequency range. To test this method we have used the strong motion data of the Geiyo earthquake of 24th March 2001, Japan recorded by KiK network. The source of this earthquake is modelled by a simple rectangular rupture of size 24 × 15 km, burried at a depth of 31 km in a multilayered earth model. This rupture plane is divided into 16 rectangular subfaults of size 6.0 × 3.75 km each. Strong motion records at eight selected near-field stations were simulated and compared with the observed records in terms of the acceleration and velocity records and their response spectrum. The comparison confirms the suitability of proposed rupture model responsible for this earthquake and the efficacy of the approach in predicting the strong motion scenario of earthquakes in the subduction zone. Using the same rupture model of the Geiyo earthquake, we compared the simulated records from our and the EGF techniques at one near-field station. The comparison shows that this technique gives records which matches in a wide frequency range and that too from simple and easily accessible parameters of burried rupture. 相似文献
10.
11.
Graeme H.McVerry 《地震工程与工程振动(英文版)》2010,9(2):177-187
A landslide displacement (DLL) attenuation model has been developed using spectral intensity and a ratio of critical acceleration coefficient to ground acceleration coefficient. In the development of the model,a New Zealand earthquake record data set with magnitudes ranging from 5.0 to 7.2 within a source distance of 175 km is used. The model can be used to carry out deterministic landslide displacement analysis,and readily extended to carry out probabilistic seismic landslide displacement analysis. DLL attenuation models have also been developed by using earthquake source terms,such as magnitude and source distance,that account for the effects of earthquake faulttype,source type,and site conditions. Sensitivity analyses show that the predicted DLL values from the new models are close to those from the Romeo model that was developed from an Italian earthquake record data set. The proposed models are also applied to an analysis of landslide displacements in the Wenchuan earthquake,and a comparison between the predicted and the observed results shows that the proposed models are reliable,and can be confidently used in mapping landslide potential. 相似文献
12.
It is demonstrated that the blind deconvolution method is fully capable of recovering the unknown Greens function and of estimating the source time functions from observed seismic data of small earthquakes. Based on the assumption of the Gaussian-mixture model of the Greens function, the newly-formulated algorithm is evaluated using synthetic seismic data along with those of the May 8, 1996 Mexico earthquake (Mc = 4.6). Since the estimated results closely match the theoretical input very well, the method is then employed to analyze the source time functions of the July 7, 1995 Pu-Li, Taiwan earthquake (ML = 5.3). The stations triggered by this event were azimuthally well covered. Using the estimated source time functions, information pertaining to the directivity effect is readily obtained, and the actual fault plane of this event is identified, thus clearly indicating that this method provides a most efficient way to estimate the source time function of a small earthquake.Acknowledgment The authors would like to express their thanks to two anonymous reviewers for their valuable suggestions and Dr. I. Santamarias courteous assistance. They also appreciate the efforts of Drs. H.C. Chiu and R.J. Rau, who provided the seismic data and the fault plane solutions. The National Science Council, Taiwan, has supported this research (NSC 91-2119-M-194-011). 相似文献
13.
The seismological model was developed initially from the fundamental relationship between earthquake ground motion properties and the seismic moment generated at the source of the earthquake. Following two decades of continuous seismological research in the United States, seismological models which realistically account for both the source and path effects on the seismic shear waves have been developed and their accuracy rigorously verified (particularly in the long and medium period ranges). An important finding from the seismological research by Atkinson and Boore and their co‐investigators is the similarity of the average frequency characteristics of seismic waves generated at the source between the seemingly very different seismic environments of Eastern and Western North America (ENA and WNA, respectively). A generic definition of the average source properties of earthquakes has therefore been postulated, referred to herein as the generic source model. Further, the generic ‘hard rock’ crustal model which is characteristic of ENA and the generic ‘rock’ crustal model characteristic of WNA have been developed to combine with the generic source model, hence enabling simulations to be made of the important path‐related modifications to ground motions arising from different types of crustal rock materials. It has been found that the anelastic contribution to whole path attenuation is consistent between the ENA and WNA models, for earthquake ground motions (response spectral velocities and displacements) in the near and medium fields, indicating that differences in the ENA and WNA motions arise principally from the other forms of path‐related modifications, namely the mid‐crust amplification and the combined effect of the upper‐crust amplification and attenuation, both of which are significant only for the generic WNA ‘rock’ earthquake ground motions. This paper aims to demonstrate the effective utilization of the latest seismological model, comprising the generic source and crustal models, to develop a response spectral attenuation model for direct engineering applications. The developed attenuation model also comprises a source factor and several crustal (wave‐path modification) component factors, and thus has also been termed herein the component attenuation model (CAM). Generic attenuation relationships in CAM, which embrace both ENA and WNA conditions, have been developed using stochastic simulations. The crustal classification of a region outside North America can be based upon regional seismological and geological information. CAM is particularly useful for areas where local strong motion data are lacking for satisfactory empirical modelling. In the companion paper entitled ‘response spectrum modelling for rock sites in low and moderate seismicity regions combining velocity, displacement and acceleration predictions’, the CAM procedure has been incorporated into a response spectrum model which can be used to effectively define the seismic hazard of bedrock sites in low and moderate seismicity regions. This paper and the companion paper constitute the basis of a long‐term objective of the authors, to develop and effectively utilize the seismological model for engineering applications worldwide. 相似文献
14.
长白山天池火山区上地壳Q值结构 总被引：4，自引：0，他引：4
同天然地震资料相比，人工地震剖面观测资料具有震源球对称且位置已知，观测点位密集、震源和观测点呈直线排列、速度结构已知等突出优点。本文提出了一种利用人工地震振幅资料反演分层介质中Q值的新方法。在利用该方法得到Q值深度函数的基础上，综合各炮点的结果，可研究Q值沿剖面的横向变化。利用上述方法计算了长白山L1剖面上地壳的Q值值结构。发现长白山天池火山区下方Q值明显低于邻近地区，该区域位于天池地区壳内低速异常体的上部，对应于已往的火山喷发通道。 相似文献
15.
为实现从注重灾后救助向注重灾前预防转变,如何提高地震灾害监测预警和风险防范能力成为我们关注的重点.本文给出了国际上GNSS位移记录、强震动加速度记录、测震速度记录在地震预警中的应用现状,并总结了各自的特点,归纳出围绕高频GNSS地震学在震级与破裂过程实时反演中的几个需要进一步研究的关键问题：（1）引入北斗系统,基于高频GNSS（GPS/BDS）双系统的实时位移解算方法来提高实时单站位移解算精度,使实时解算精度达到厘米级;（2）开展强震仪加速度记录基线偏移校正研究,弥补地震近场GNSS站密度不足问题;（3）强震仪加速度记录与GNSS位移记录特点不同,开展强震仪加速度数据与GNSS位移数据实时融合处理研究,快速获得包含丰富地震形变和速率的波形数据;（4）测震学方法可快速估算震级,但是在强震发生时会出现震级饱和现象,造成震级估算偏低.需要开展基于GNSS位移时间序列的多种方法相结合的实时震级估算方法研究,通过与地震学方法比较和结合,来得到精度高、计算快的震级估值算法;（5）基于高频GNSS、断层初始模型快速选取、断层尺度、参数自适应调整是快速判断断层破裂方向的基础,在断层破裂过程自适应准实时反演算法方面需要进一步加强.通过国内外研究现状调研、分析,表明基于高频GNSS地震学的震级快速确定、震源破裂过程准实时反演算法的发展将对我国地震预警系统从"二网融合"到"三网融合"提供坚实的技术支撑. 相似文献
16.
Vladimir Yu. Sokolov Chin-Hsuing Loh Kuo-Liang Wen 《Soil Dynamics and Earthquake Engineering》2003,23(8):715-735
We analyze the ability of different spectral models to describe the frequency content of ground motion during the 1999 Chi–Chi earthquake (MW=7.6, Taiwan) and two large (ML=6.8) aftershocks. The spectral models evaluated include the one-corner model of Brune applied with various key parameters (seismic moment and stress drop), and the two-corner-frequency models proposed for eastern North America [Bull. Seismol. Soc. Am. 83 (1993) 1778] and California [Bull. Seismol. Soc. Am. 90 (2000) 255]. The ground-motion spectra predicted by these spectral models for hypothetical very hard rock site were compared with the Chi–Chi earthquake data obtained on rock (class B) and soft rock or very dense soil (class C) sites. The approach also allows us evaluating the generalized empirical amplification function for class B and C sites in the region.
It has been found that, the amplitude spectra of recorded ground acceleration (the mainshock and aftershocks) for frequencies larger than 0.3–0.4 Hz agree with the modelled two-corner-frequency spectra calculated using the model proposed for California. The single-corner-frequency model also provides a good agreement with the observations when using so-called ‘short-period seismic moment’ [Phys. Earth Planet. Interiors 37 (1985) 108] instead of the reported values obtained from long-period waves. The key parameters used in the single-corner model coincide with parameters of subsources evaluated for the complicated mainshock source. Therefore, it is possible to confirm the suggestion that the short-period seismic waves, at least for the thrust earthquakes, are generated mainly from the fracture of small-scale heterogeneities. The use of two-corner-frequency source model for earthquake spectrum that is based on long-period seismic moment value is equivalent, for frequencies larger than 0.3–0.4 Hz, to the use of single-corner-frequency model that is based on the parameters of major subsource. 相似文献
17.
Another look at the 1993 and 1995 Gulf of Aqaba earthquakes from the analysis of teleseismic waveforms 总被引：3，自引：0，他引：3
In this work we estimated the source process and the source parameters of the 1993 and 1995 Gulf of Aqaba earthquakes. To
investigate the source process of both events, we inverted the teleseismic P- and SH-waveform data using an iterative technique in which the rupture is modeled as a series of point source subevents with varying
mechanisms. The main source process of the 1993 event can be explained by two subevents with essentially the same mechanism.
These two subevents had different focal depths. The second subevent was about 5 s later than the first one and about 70% of
the moment rate released with the second one. The total rupture duration time was about 12 s. Our solution for the 1995 Gulf
of Aqaba earthquake indicated that the event consists of three subevents with various fault geometries and about 60% of the
moment released by the second subevent. The first subevent occurred at a depth of 13 km, and was followed after 9 s by the
second one at a depth of 11 km; the third subevent, at a depth of 9 km occurred 19 s after the second one. The total duration
of the earthquake rupture process was about 18.7 s. The obtained mechanisms for the 1993 and 1995 Gulf of Aqaba earthquakes
are well correlated with the structural setting of the Gulf of Aqaba. 相似文献
18.
本文介绍了2015年4月25日尼泊尔Mw7.9(MS8.1)地震发生后的破裂过程快速反演工作,以及后续开展的地震波与少量GPS资料的初步联合反演工作.两项工作得到的反演结果尽管在最大滑动量估计方面存在一些差别,但都一致地显示此次地震是发生在低倾角俯冲断裂上的一次单侧破裂事件,破裂主要朝东南方向传播;断层滑动主要发生在震中至加德满都一带.在加德满都附近区域,其下方破裂与朝东南传播的地震波的多普勒聚焦效应可能造成较强的震感和较大的破坏.对比历史大地震发现,2015年尼泊尔Mw7.9地震的浅部破裂紧邻1934年Mw8.2地震的地表破裂,余震分布与1833年M7.6地震的宏观震中基本重合,其破裂填补了前两次地震破裂以西100km左右的空区,表明此次地震是1934年Mw8.2地震与1833年M7.6地震向西继续延伸的结果. 相似文献
19.
Large magnitude earthquakes generated at source–site distances exceeding 100km are typified by low‐frequency (long‐period) seismic waves. Such induced ground shaking can be disproportionately destructive due to its high displacement, and possibly high velocity, shaking characteristics. Distant earthquakes represent a potentially significant safety hazard in certain low and moderate seismic regions where seismic activity is governed by major distant sources as opposed to nearby (regional) background sources. Examples are parts of the Indian sub‐continent, Eastern China and Indo‐China. The majority of ground motion attenuation relationships currently available for applications in active seismic regions may not be suitable for handling long‐distance attenuation, since the significance of distant earthquakes is mainly confined to certain low to moderate seismicity regions. Thus, the effects of distant earthquakes are often not accurately represented by conventional empirical models which were typically developed from curve‐fitting earthquake strong‐motion data from active seismic regions. Numerous well‐known existing attenuation relationships are evaluated in this paper, to highlight their limitations in long‐distance applications. In contrast, basic seismological parameters such as the Quality factor (Q‐factor) could provide a far more accurate representation for the distant attenuation behaviour of a region, but such information is seldom used by engineers in any direct manner. The aim of this paper is to develop a set of relationships that provide a convenient link between the seismological Q‐factor (amongst other factors) and response spectrum attenuation. The use of Q as an input parameter to the proposed model enables valuable local seismological information to be incorporated directly into response spectrum predictions. The application of this new modelling approach is demonstrated by examples based on the Chi‐Chi earthquake (Taiwan and South China), Gujarat earthquake (Northwest India), Nisqually earthquake (region surrounding Seattle) and Sumatran‐fault earthquake (recorded in Singapore). Field recordings have been obtained from these events for comparison with the proposed model. The accuracy of the stochastic simulations and the regression analysis have been confirmed by comparisons between the model calculations and the actual field observations. It is emphasized that obtaining representative estimates for Q for input into the model is equally important.Thus, this paper forms part of the long‐term objective of the authors to develop more effective communications across the engineering and seismological disciplines. Copyright © 2003 John Wiley & Sons, Ltd. 相似文献
20.
建筑物受损信息是地震受灾程度评估的基础,针对传统建筑物表面信息识别人工成本高、效率低等问题,受深度学习提取建筑物影像的启发,提出利用无人机倾斜摄影模型与深度学习相结合的方法提取震后建筑物表面破损信息。以2019年长宁6.0级地震为例,选用双河镇震后倾斜摄影模型切片图为数据源,对比分析面向对象分类方法、VGG-16模型和DeeplabV3+模型对建筑物表面损毁信息的提取结果。分析结果表明,针对建筑物表面破损信息的提取,尤其是细小裂缝的提取,语义分割网络DeeplabV3+模型具有较强的优势(准确率96.93%、召回率96.85%、总体精度96.89%),可实现建筑物表面破损信息的有效提取,具有较强的应用价值。 相似文献