一种多层特征融合的道路场景实时分割方法     

周继苗  李必军  陈世增 《测绘通报》2020,(1):10-15

道路场景理解是自动驾驶领域中重要模块之一，它可以提供关于道路更丰富的信息，对于建立高精度地图和实时规划都具有重要作用。其中，语义分割可以为图像每个像素赋予类别信息，是自动驾驶场景理解中最常用的方法。但是，目前常用的语义分割算法在速度和精度上大都不能达到很好的平衡。本文在MobileNetV2的基础上，提出了一种多层次特征融合的方法，使得网络可以在实时运行的同时保证精度满足实际应用的需求，并在Cityscapes数据集上进行了试验验证和分析。  相似文献   

高分辨率遥感影像多尺度均值漂移分割算法研究   总被引：4，自引：3，他引：1  

沈占锋  骆剑承  胡晓东  孙卫刚 《武汉大学学报(信息科学版)》2010,(3)

根据高分辨率遥感影像信息提取过程中对影像的对象化分割的需求,分析了均值漂移分割算法的原理,并对其多尺度分割方法进行了设计与实现。实验证明该算法具有较好的影像分割精度。  相似文献   

An improved segmentation approach for planar surfaces from unstructured 3D point clouds     

Tarek M.  Awwad  Qing  Zhu  Zhiqiang  Du  Yeting  Zhang 《The Photogrammetric Record》2010,25(129):5-23

The extraction of object features from massive unstructured point clouds with different local densities, especially in the presence of random noisy points, is not a trivial task even if that feature is a planar surface. Segmentation is the most important step in the feature extraction process. In practice, most segmentation approaches use geometrical information to segment the 3D point cloud. The features generally include the position of each point (X, Y and Z), locally estimated surface normals and residuals of best fitting surfaces; however, these features could be affected by noisy points and in consequence directly affect the segmentation results. Therefore, massive unstructured and noisy point clouds also lead to bad segmentation (over-segmentation, under-segmentation or no segmentation). While the RANSAC (random sample consensus) algorithm is effective in the presence of noise and outliers, it has two significant disadvantages, namely, its efficiency and the fact that the plane detected by RANSAC may not necessarily belong to the same object surface; that is, spurious surfaces may appear, especially in the case of parallel-gradual planar surfaces such as stairs. The innovative idea proposed in this paper is a modification for the RANSAC algorithm called Seq-NV-RANSAC. This algorithm checks the normal vector (NV) between the existing point clouds and the hypothesised RANSAC plane, which is created by three random points, under an intuitive threshold value. After extracting the first plane, this process is repeated sequentially (Seq) and automatically, until no planar surfaces can be extracted from the remaining points under the existing threshold value. This prevents the extraction of spurious surfaces, brings an improvement in quality to the computed attributes and increases the degree of automation of surface extraction. Thus the best fit is achieved for the real existing surfaces.  相似文献   

Modeling three‐dimensional hydraulic fracture propagation using virtual multidimensional internal bonds     

K. Huang  Z. Zhang  Ahmad Ghassemi 《国际地质力学数值与分析法杂志》2013,37(13):2021-2038

Propagation of fractures, especially those emanating from wellbores and closed natural fractures, often involves Mode I and Mode II, and at times Mode III, posing significant challenges to its numerical simulation. When an embedded inclined fracture is subjected to compression, the fracture edge is constrained by the surrounding materials so that its true propagation pattern cannot be simulated by 2D models. In this article, a virtual multidimensional internal bond (VMIB) model is presented to simulate three‐dimensional (3D) fracture propagation. The VMIB model bridges the processes of macro fracture and micro bond rupture. The macro 3D constitutive relation in VMIB is derived from the 1D bond in the micro scale and is implemented in a 3D finite element method. To represent the contact and friction between fracture surfaces, a 3D element partition method is employed. The model is applied to simulate fracture propagation and coalescence in typical laboratory experiments and is used to analyze the propagation of an embedded fracture. Simulation results for single and multiple fractures illustrate 3D features of the tensile and compressive fracture propagation, especially the propagation of a Mode III fracture. The results match well with the experimental observation, suggesting that the presented method can capture the main features of 3D fracture propagation and coalescence. Moreover, by developing an algorithm for applying pressure on the fracture surfaces, propagation of a natural fracture is also simulated. The result illustrates an interesting and important phenomenon of Mode III fracture propagation, namely the fracture front segmentation. Copyright © 2012 John Wiley & Sons, Ltd.  相似文献   

A multiscale finite element formulation for axisymmetric elastoplasticity with volumetric locking     

Kaiming Xia  Kenn K. Q. Zhang 《国际地质力学数值与分析法杂志》2010,34(10):1076-1100

Similar to plane strain, axisymmetric stress problem is also highly kinematics constrained. Standard displacement‐based finite element exhibits volumetric locking issue in simulating nearly/fully incompressible material or isochoric plasticity under axisymmetric loading conditions, which severely underestimates the deformation and overestimates the bearing capacity for structural/geotechnical engineering problems. The aim of this paper is to apply variational multiscale method to produce a stabilized mixed displacement–pressure formulation, which can effectively alleviate the volumetric locking issue for axisymmetric stress problem. Both nearly incompressible elasticity and isochoric J₂ elastoplasticity are investigated. First‐order 3‐node triangular and 4‐node quadrilateral elements are tested for locking issues. Several representative simulations are provided to demonstrate the performance of the linear elements, which include the convergence study and comparison with closed‐form solutions. A comparative study with pressure Laplacian stabilized formulation is also presented. Copyright © 2009 John Wiley & Sons, Ltd.  相似文献   

基于粒子群优化算法的最大类间方差多阈值图像分割     

周晓伟  葛永慧 《测绘科学》2010,35(2):88-89,122

最大类间方差法是图像分割中一种常用的阈值分割方法,对于单阈值分割具有显著的效果,但是对于多阈值分割,计算复杂度大、耗时较多。本文将粒子群优化算法与最大类间方差法结合,提出了一种新的图像分割方法,该方法利用粒子群优化算法的寻优高效性,并由灰度图像的最大类间方差值作为适应值,搜索最优分割阈值,实现图像的多阈值分割。实验结果显示,新方法大大缩短了寻找最优阈值的时间,降低了运算复杂度,提高了图像分割速度,说明基于粒子群优化算法的图像分割算法是可行的、有效的。  相似文献   

基于LIDAR数据的建筑物轮廓提取   总被引：2，自引：0，他引：2  

杨洋  张永生  马一薇  胥亚 《测绘科学》2010,35(3):203-205

建筑物轮廓的准确提取是建筑物三维重建中最重要的一步。本文在研究已有建筑物轮廓提取方法的基础上,针对LIDAR离散的点云数据,提出了一种自动快速提取建筑物轮廓信息的方法。首先通过点云数据生成城市的数字表面模型(DSM)和数字地面模型(DTM)相减计算得出规则化的数字表面模型(nDSM),进而将地面点和非地面点进行分类;其次,考虑到地物的几何特性,提出一种8邻域搜索的方法对非地面点点云进行分割,得到建筑物表面点云;最后运用基于梯度图的边界跟踪的方法来获取建筑物的轮廓信息。实验表明:该方法能有效地提取建筑物轮廓。  相似文献   

地理编码系统中地名地址分词算法研究   总被引：4，自引：0，他引：4  

张林曼  吴升 《测绘科学》2010,35(2):46-48

对中文分词有关技术进行分析和研究,结合地理编码的特点,设计了基于双字哈希和数组的三层组合分词词典数据结构,以及逐次增字的最大正向分词算法。并通过构建和遍历地名通名词典,识别地名地址中未登录词。测试表明,本文算法速度和效果较好,解决了地理信息系统中中文分词的相关问题。  相似文献   

纹理主方向的遥感影像居民地提取   总被引：1，自引：0，他引：1  

金飞  张占睦  芮杰 《测绘科学》2010,35(4):139-141

本文研究利用基于傅立叶变换和Hough变换的灰度共生矩阵方法对遥感影像居民地进行提取。首先,通过傅立叶变换把遥感影像变换到频率域,然后利用Hough变换找出遥感影像的主纹理方向,进而把影像主纹理方向旋转至水平,最后利用灰度共生矩阵方法进行纹理分割。实验结果表明:基于傅立叶变换和Hough变换的灰度共生矩阵方法对遥感影像居民地有较好的提取结果。  相似文献   

一次自上向下发展的高原涡的多尺度动力学分析     

容逸能  马继望  李瑶婷  梁湘三 《气象科学进展》2021,(1)

高原涡作为经常给我国带来暴雨等灾害的天气系统,其形成一般认为是通过感热和潜热自下而上激发的,然而,2013年5月下旬发生的一次引发其下游灾害性强降水的高原涡却是由对流层高层天气尺度低涡诱发的。为此,基于新发展的多尺度子空间变换和多尺度能量涡度方法以及ERA5再分析资料对其动力学过程进行了详尽的探讨,先将原始场重构到三个尺度子空间,即背景环流尺度子空间、天气尺度子空间和高频尺度子空间,重构场上首次显示此次过程生成于青藏高原西北侧,其成因为对流层高层基本气流尺度向天气尺度的跨尺度动能正则传输,即正压失稳,并且表现为从高层向下。在发展阶段,其能量最终来源为基本气流向天气尺度的有效位能传输和非绝热加热,然而这些过程只发生于涡旋低层的西侧。进一步分析发现,天气尺度内存在一个能量再分配“路径”:首先,低层西侧获得的有效位能转换为动能,西侧垂直的气压梯度力做功将低层获得动能向高层分配;在高层,水平的气压梯度力做功进而将西侧获得的动能向东侧分配;东侧垂直的气压梯度力做功再将动能向低层分配;至此,低层西侧获得的能量被分配到整个涡旋空间中,使得涡旋能够均匀发展。  相似文献