基于深度学习的公路施工宏观状态快速评估

高速公路是重要的基础建设项目，针对传统公路施工状态评估效率低下问题，本文提出一种基于深度学习的施工状态快速评估方法。选取高分辨率无人机影像进行实验，首先根据施工状态对目标进行标注与切分，制作深度学习数据集，输入DeepLabV3深度学习网络中进行训练，构建出施工状态分类模型，进而对研究区进行分类。两个研究区的实验证明，本文方法状态识别精度优于同类方法，基于本文方法能够实现公路施工宏观状态准确快速评估。     

基于深度学习的公路隧道表观病害智能识别研究现状与展望

公路隧道服役过程中会产生诸多衬砌病害,其会影响隧道的结构耐久性与运营安全性,对隧道表观病害进行高效智能化识别至关重要。常用的人工巡检方式效率低下且准确率低,而基于深度学习算法进行表观病害智能识别能提高检测的效率和准确性,相较于传统方法而言在实际隧道工程中具有更好的应用前景。利用深度学习可以学习隧道病害的特征信息,有利于未来隧道病害识别智能化的发展。简述深度学习在隧道表观病害识别中的应用原理,从人工拍照方法、数字图像采集和激光扫描技术三方面介绍病害图像的采集,从标注软件和数据增强方法总结数据集的构建和扩充方法,在图像分类、目标检测、语义分割三方面总结深度学习算法在隧道病害检测的应用现状,归纳当前应用的不足之处,最后分析与展望深度学习在隧道表观病害智能化识别方向广泛应用需要研究的问题与方向。     

基于残差网络的混凝土结构病害分类识别研究

针对混凝土结构病害识别类型单一、精度较低的现状,提出了基于残差网络和迁移学习的病害分类识别方法,通过构建多属性病害数据集,利用迁移学习优化残差网络模型,提出混凝土结构健康状态识别的多个任务。首先收集混凝土结构的病害状态图像,依次通过数据清洗、尺寸均一化、数据扩增和多人投票标注,最终得到包含6 680张图像的混凝土结构病害多属性数据集,并依据不同标注属性进行了相应训练集、验证集和测试集的划分; 然后利用迁移学习对预训练的ResNet-34网络前3个部分进行参数冻结,后续2个部分的参数进行重新训练,并在模型末端添加新的参数,基于已构建的数据集进行训练; 最后在提出的构件类别检测、剥落检测、病害检测和病害类别检测任务中,分别获得84.88%、98.56%、97.18%和85.34%的F1分数。结果表明:通过构建多属性标注的混凝土结构病害数据集训练深度学习模型,可较好地实现多场景特征下的病害识别效果; 采用迁移学习技术可从开源数据中获取较好的特征提取效果; 改进的ResNet-34网络可克服网络退化问题,并针对混凝土结构病害识别的多个任务获得较好的效果; 相对于单一的混凝土结构病害识别,进行病害部位、程度、多类别的系统性检测,可为结构状态评估提供详细信息,更贴合工程实际需要。     

人工智能辅助设计：基于深度学习的 风景园林平面识别与渲染

采用风景园林学与人工智能的跨学科研究方式,开发了一种将深度学习模型——生成对抗网络(Generative Adversarial Network,GAN)用于风景园林平面图用地识别与图像渲染的新应用场景。以325张细致标注的平面方案图建立用于深度学习的数据集,训练循环生成对抗网络(CycleGAN)实现平面图不同用地类型地块的提取任务,以及平面色块图到色彩肌理图的渲染生成。进一步从图片质量、正确规范性和色彩表达等方面评价模型的识别与渲染结果。该训练模型有潜力被应用于风景园林案例的用地类型分析及平面渲染,帮助设计师提升分析及制图效率。     

基于合成点云与深度学习的建筑构件智能识别方法

本文研究了利用三维建筑信息模型生成的合成点云来训练深度学习算法以实现建筑构件智能识别的可行性。为了实现这一目标,本文首先提出了一种通过三种常见的商业软件将建筑信息模型转换为合成点云的原始方法。然后使用这些合成点云作为模拟数据集来训练深度学习模型,比较在不同数据集(真实数据集与合成数据集)下训练模型的智能识别性能,以验证合成点云数据集的有效性。实验结果证明了利用建筑信息模型生成的合成点云实现智能识别的可行性,合成数据集与真实数据集的训练模型其识别准确率仅相差3%,进一步表明了在智能识别中使用合成数据集代替真实数据集的可能性。该方法也为研究人员提供了一种新的方法来构建特定的数据集,用于他们自己的智能识别与语义分割研究,并为三维重建工作做出了贡献。     

Context Capture实景三维建模硬件性能分析

基于图像自动匹配技术的实景三维建模,近年来得到广泛应用,该类建模主要包含连接点提取、影像对选择、空间方位定向、连接点匹配、光束法平差、构建不规则三角网、三角网的优化与光滑,纹理映射等关键步骤。整个流程中对计算机硬件的要求各异,部分流程需要多核多线程,部分流程以单核运算为主。如何最大限度的利用计算机资源,完成实景三维建模工作值得思考。文章从实景三维建模关键技术流程和硬件主要指标进行简要的分析,分析内容主要包含中央处理器、GPU与Vulkan、内存、磁盘性能、网络、分布式集群。     

基于AutoLISP的三维模型快速建立与精度分析

建筑物的三维模型是数字城市重建的重要组成部分，而建筑物高精度三维模型的快速建立一直是三维数字仿真亟待解决的问题。本文通过无协作目标电子全站仪采集建筑的三维坐标，基于AutoLISP实现了数据的导入、建筑物墙面的拟合、特征点的投影及门窗等构件的自动生成，从而快速生成建筑物的三维模型，并对生成的建筑模型进行精度分析。基于AutoLISP构建建筑物的三维模型，大大提高了建模速度；并能有效控制三维建筑模型的质量。     

基于Pro/E的液压挖掘机工作装置运动仿真

近年来,集挖掘和装载功能于一体的液压挖掘机在基础建设和民用建筑建设中的使用与日俱增。由于液压挖掘机的工作条件比较恶劣,工作装置故障较多,造成整机工作可靠性较差,因此挖掘机的可靠性和最优化设计成为国产挖掘机设计的重点和难点。本文以某国产22t挖掘机为例,探讨在Pro/E软件环境下,对挖掘机工作装置进行三维实体建模、虚拟装配、运动仿真与动态模拟,为挖掘机物理样机的制造、新机型设计方案的评估提供有效参考数据。1挖掘机虚拟样机的建立1·1挖掘机三维零件模型的建立一般来说,在Pro/E软件环境下,机械系统的三维建模应该严格按照…     

基于深度学习的三维点云自动生成BIM模型方法

在数字孪生、智慧城市的浪潮下,建筑行业正积极探索一种能快速重塑"既有建筑"3D信息模型的方法.从深度学习3D目标检测算法角度出发,着手大规模建筑数据集的生成和点云深度学习理论,分析点云深度学习框架所需输入数据类型,重点介绍了各类建筑构件的3D边界框及三维点云的创建过程,对比具有相同数据结构的不同点云数据集并实现了基于S...     

基于MAXScript的城市综合管网批量建模方法

针对城市综合管网三维建模过程中手工建模方法效率低,而完全从底层开发或基于商业软件进行二次开发实现管网建模具有成本高、周期长及成果可移植性差等缺点,提出一种在3ds Max软件环境下,基于MAXScript脚本语言进行管线和附属设施批量建模的技术方法,实现了城市综合管网的整体自动建模,所建模型可转换为OSG、3DS等通用的三维数据交换格式,可方便地实现数据共享和重复利用。最后对该建模方法进行了生产实例验证,并将生成的三维管网模型导入Arc GIS平台进行属性挂接和数据管理,结果表明该方法简单、可靠,在一定程度上节约了三维管网数据建设的时间和成本。     

A 3D model based control of an excavator

New technologies facilitate machine control systems, which use data from CAD and 3D measurement systems to automatically control construction machinery, such as an excavator. This paper examines the possibilities of controlling a six degrees of freedom (DOF) excavator with the final objective of controlling the movements of the excavator by using a positioning system such as a GPS in conjunction with a CAD model of the road surface. In comparison with the traditional 4 DOF machine, the excavator was provided with two additional degrees of freedom by applying a 2 DOF Rototilt, an accessory commonly in use. To study this problem, Msc.Adams with Matlab/Simulink was used as a simulation environment. In this paper the path of the bucket is calculated by using the CAD reference of the target plane.     

A large-scale dataset of buildings and construction sites

With the rapid development of deep learning and machine automation technology, as well as workforce aging, increasing labor costs, and other issues, an increasing number of scholars have paid attention to the use of these techniques to solve problems in civil engineering. Although progress has been made in applying deep learning to damage detection, many subfields in civil engineering are still in the initial stage, and a large amount of data has not been used. Moreover, the rapid development of a field cannot be separated from large open-source datasets and many researchers. Therefore, this study attempts to construct a dataset named the BCS dataset of nearly 212,000 photos of buildings and construction sites using multi-threaded parallel crawler technology and offline collection. The dataset will be expanded regularly. As a practical demonstration, the StyleGAN3 and StyleGAN2 generative adversarial networks were utilized on the dataset to create faked safety hat images and high-resolution architectural images. Subsequently, four classic classification models were employed to validate the dataset, achieving a Top-1 accuracy of up to 0.947. These results underscore the dataset's excellent potential for practical applications.     

Automated annotation for visual recognition of construction resources using synthetic images

The recognition of construction equipment is always necessary and important to monitor the progress and the safety of a construction project. Recently, the potentials of computer vision (CV) techniques have been investigated to facilitate the current equipment recognition method. However, the process of manually collecting and annotating a large image dataset of different equipment is one of the most time-consuming tasks that may delay the application of the CV techniques for construction equipment recognition. Moreover, collecting effective negative samples brings more difficulties for training the object detectors. This research aims to introduce an automated method for creating and annotating synthetic images of construction equipment while significantly reducing the required time. The synthetic images of the equipment are created from the three-dimensional (3D) models of construction machines combined with various background images taken from construction sites. The location of the equipment in the images is known since that equipment is the only object over the single-color background. This location can be extracted by applying segmentation techniques and then used for the annotation purpose. Furthermore, an automated negative image sampler is introduced in this paper to automatically generate many negative samples with different sizes out of one general image of a construction site in a way that the samples do not include the target object. The test results show that the proposed method is able to reduce the required time for annotating the images in comparison with traditional annotation methods while improving the detection accuracy.     

液压挖掘机运动学问题及可控性与可观性分析

本文以哈尔滨工业大学工程机械实验室液压挖掘机样机为模型，阐述了液压挖掘机实现自动控制的方法；提出了运动学正问题、逆问题及其解法；针对样机模型的液压控制系统数学模型，从现代控制理论的角度，对其可控性和可观性进行分析，并进行了仿真。     

Image dataset development for measuring construction equipment recognition performance

The recognition of construction operational resources (equipment, workers, materials, etc.) has played an important role in achieving fully automated construction. So far, many object recognition methods have been developed in computer vision; however, they have been tested with a few categories of objects in natural scenes. Therefore, their performance on the recognition of construction operational resources is unclear, especially considering construction sites are typically dirty, disorderly, and cluttered. This paper proposes a standard dataset of construction site images to measure the construction equipment recognition performance of existing object recognition methods. Thousands of images have been collected and compiled, which cover 5 classes of construction equipment (excavator, loader, dozer, roller and backhoe). Each image has been annotated with the equipment type, location, orientation, occlusion, and labeling of equipment components (bucket, stick, boom etc.). The effectiveness of the dataset has been evaluated with two well-known object recognition methods in computer vision. The results show that the dataset could successfully identify the performance of these methods in terms of correctness, robustness, and speed of recognizing construction equipment.     

3D AR-based modeling for discrete-event simulation of transport operations in construction

This study proposes a 3D visualized modeling method for DES of transport operations in construction. The 3D simulation model built is a virtual field scene with property settings. AR technology was further applied to allow the use of a real-world image as the modeling background, which pictorially presents the current status of the real site as a visual basis for modeling. A typical transport operation was analyzed to determine the component classes for modeling. Then the visual representation and attributes of each modeling component class were proposed, along with modeling rules to build the 3D simulation model. A prototype system with STROBOSCOPE as the simulation engine was developed for presenting the proposed modeling method. A set of transformation rules was proposed for converting a 3D simulation model to a STROBOSCOPE input file. The system automatically extracts the simulation output and animates the 3D model to visually demonstrate the simulation result.     

Deep domain adaptation based video smoke detection using synthetic smoke images

In this paper, a deep domain adaptation based method for video smoke detection is proposed to extract a powerful feature representation of smoke. Due to the smoke image samples limited in scale and diversity for deep CNN training, we systematically produced adequate synthetic smoke images with a wide variation in the smoke shape, background and lighting conditions. Considering that the appearance gap (dataset bias) between synthetic and real smoke images degrades significantly the performance of the trained model on the test set composed fully of real images, we build deep architectures based on domain adaptation to confuse the distributions of features extracted from synthetic and real smoke images. This approach expands the domain-invariant feature space for smoke image samples. With their approximate feature distribution separated from non-smoke images, the recognition rate of the trained model is improved significantly compared with the model trained directly on mixed dataset of synthetic and real images. Experimentally, several deep architectures with different design choices are applied to the smoke detector. The ultimate framework can get a satisfactory result on the test set. We believe that our method own strong robustness and may offer a new way for the video smoke detection.     

议施工阶段的造价管理

从材料及机械设备管理、工程进度款管理、工程变更管理、工程索赔管理四方面入手,分析研究了施工阶段的造价管理要点,以期通过科学的管理方法实现建设单位的造价控制目标。     

无人机技术在工程建设领域的应用研究

建设领域的应用现状,在梳理无人机技术基本原理、“无人机性能研究?技术初步交叉融合?技术广泛交叉融合”3 个发展阶段的基础上,分析无人机技术在工程项目不同阶段的应用现状及未来应用前景,无人机是很好的自动化采集工具,与三维点云技术密不可分,通过三维影像技术在厦门地铁三号线车站大型施工场地的应用实践,对无人机辅助工程项目施工阶段的场地布置和进度管控中的技术可行性进行探讨,并提出有效利用无人机开展三维建模工作的建议。