Infrastructure monitoring is being transformed by the advancements on remote sensing, unmanned vehicles and information technology. The wide interaction among these fields and the availability of reliable commercial technology are helping pioneer intelligent inspection methods based on digital 3D models. Commercially available Unmanned Aerial Vehicles (UAVs) have been used to create 3D photogrammetric models of industrial equipment. However, the level of automation of these missions remains low. Limited flight time, wireless transfer of large files and the lack of algorithms to guide a UAV through unknown environments are some of the factors that constraint fully automated UAV inspections. This work demonstrates the use of unsupervised Machine Learning methods to develop an algorithm capable of constructing a 3D model of an unknown environment in an autonomous iterative way. The capabilities of this novel approach are tested in a field study, where a municipal water tank is mapped to a level of resolution comparable to that of manual missions by experienced engineers but using $63\%$ . The iterative approach also shows improvements in autonomy and model coverage when compared to reproducible automated flights. Additionally, the use of this algorithm for different terrains is explored through simulation software, exposing the effectiveness of the automated iterative approach in other applications.

The smart society, including smart infrastructures is developing quickly. The smart monitoring of the water distribution systems is a significant part of this development, trying to address possible problems in the Water Distribution Networks (WDNs) such as water leakages, pressure instability and water contamination. By sampling and sensing important parameters in the water distribution infrastructure and sending these data to control systems through a low power wide area network (LPWAN), a Cyber Physical System (CPS) of a digitalized WDN can be built. This thesis provides support for the design of such a CPS, by the design of a simulation framework, which includes coordinated WDN and communication network simulators. First, we make a selection on the communication network simulators and protocol stacks. Network simulator 3 (NS-3) is chosen as the simulation tool and Long Range (LoRa)/long range wide-area network (LoRaWAN) is chosen as the network protocol. The EPANET software is used as WDN simulator. Then, the existing implementation of the LoRaWAN protocol stack in NS-3 is modified, to allow the network simulator run parallel to the WDN simulator, and to transmit WDN sampling data to a control center. Finally, the effect of the communication network properties on the performance of the WDN monitoring system is evaluated, and it is shown that the communication delays can affect the monitoring performance even in small systems. The thesis provides the first steps of the development of a simulation environment of the cyber physical system of digitalized water distribution networks, and is expected to support further research in the area. / Det smarta samhället, inklusive smarta infrastrukturer, utvecklas snabbt. Den smarta övervakningen av vattendistributionssystemen är en betydande del av denna utveckling, som försöker lösa eventuella problem i vattendistributionsnäten (WDN) såsom vattenläckage, tryckinstabilitet och vattenförorening. Genom att ta prov och känna av viktiga parametrar i vattendistributionsinfrastrukturen och genom att skicka dessa data till ett kontrollsystem via ett LPWAN, kan ett CPS av ett digitaliserat WDN byggas. Detta examensarbete ger stöd för designen av en sådan CPS, genom utformningen av ett simuleringsramverk, som inkluderar koordinerade simulatorer för WDN och kommunikationsnätverk. Först gör vi ett urval på simulatorer för protokollstackar och kommunikationsnätverk. NS-3 väljs som simuleringsverktyg och LoRa/LoRaWAN väljs som nätverksprotokoll. EPANET-mjukvaran används som WDN-simulator. Sedan modifieras den befintliga implementeringen av LoRaWAN-protokollstacken i NS-3, för att tillåta nätverkssimulatorn att köras parallellt med WDN-simulatorn och för att överföra WDN-samplingsdata till ett kontrollcenter. Slutligen utvärderas effekten av kommunikationsnätverkets egenskaper på prestandan hos WDN-övervakningssystemet, och det visas att förseningarna i kommunikationsnätet kan påverka övervakningsprestandan även i små system. Avhandlingen ger de första stegen i utvecklingen av en simuleringsmiljö av det cyberfysiska systemet för digitaliserade vattendistributionsnät, och förväntas stödja ytterligare forskning inom området.

<p dir="ltr">The concept of Intelligent Transport Systems (ITS) for Smarter Cities has been introduced to categorize advanced tools and solutions to facilitate intelligent and data-driven monitoring and management of the transportation system. Improvements in visual data collection and recognition offer promising potential for low-cost highway infrastructure health monitoring with mobility and non-destructiveness. However, innovative computer vision-based techniques take place against a backdrop of deteriorating highway systems. Traditional and current state of practices in the area continue applications involving laborious and time-consuming surveys conducted by trained human inspectors or high-cost sensor networks with limited performance and/or unsolved practical challenges. Moreover, the ASCE <i>2021 Infrastructure Report Card</i> (ASCE, 2021) diagnosed that over 40% of public roads are in poor or mediocre condition and the increasing needs of enhancement and expansion for the system has result in a massive 786 billion USD of underfunding. Hence, adequate and low-cost monitoring and assessment methods for traffic and pavement are indispensable for bringing an optimal life-cycle operations and maintenance management of highway infrastructures. A highway pavement is expected to withstand ever-increasing numbers of heavy vehicles of different types. With the introduction of autonomous vehicles, the traffic loading impact is further increasing. As such, accurate vehicle counts with classification and load identification is a vital traffic monitoring task. Moreover, period assessment of pavement friction performance at the tire-road interaction has a pivotal role in developing the inventory of the condition of infrastructures at the highway network level.</p><p dir="ltr">This research seeks to improve the measurement quality and monitoring efficiency of current traffic monitoring and pavement friction assessment practices, taking the use of modern digital techniques and computer vision advances. To achieve this objective, the study is realized through three research procedures: (i) development of a vision-based highway traffic load monitoring method towards WIM quality control, (ii) a mechanistic-empirical investigation of locked wheel friction testing on highway horizontal curves, and (iii) an outdoor photogrammetric survey of pavement surface with texture roughness characterization towards accurate and contactless friction prediction.</p><p dir="ltr">The first task, the vision-based traffic monitoring method uses convolutional neural network (CNN)-based semantic segmentation and object tracking algorithms. This approach began with pixel-wise recognition and localization of vehicle axles and body components using a real-world traffic image. Based on parameters estimated by a target-free camera calibration, the recognition results are used to identify the vehicle in image at a fine-grained classification level. Finally, according to the vehicle time and location synchronization, the camera machine vision automates quality control of a local Weigh-in-Motion (WIM) system through information fusion.</p><p dir="ltr">In the second task a locked wheel skid tester (LWST) practicability on highway horizontal curves is investigated through installation of two consumer-grade action cameras. Simultaneous monitoring of LWST test tire sideslip angle and vehicle articulation angle are achieved using image processing techniques such as feature detector and optical flow. The vehicle testing speed and pavement geometric properties and their impacts on pavement friction test performance are evaluated through field experiments under distinct testing scenarios. In the end, data analysis is performed to discuss the test reliability on both theoretical and practical grounds.</p><p dir="ltr">The third task develops a synthetic environment for outdoor practices of high-quality 3D pavement texture scan using Structure-from-Motion. A photo-realistic computer graphics model of asphalt pavement surface is produced and virtually scanned using candidate image acquisition plans. Quality assessment of the corresponding 3D point cloud reconstruction models is performed to suggest an optimal pavement texture sampling rate for field measurements. Then, a field texture photogrammetric survey is conducted, followed by measurements of a comprehensive list of texture parameters using pavement surface reconstruction sampled at different spatial resolution levels over the micro-texture scales. Finally, three LWST friction number prediction models are implemented using the texture characteristics as predicators. And the model with best test performance is proposed with a conclusion of working range of texture sampling rate.</p>

Structural health monitoring and building assessment are crucial to acquire structures’ states and maintain their conditions. Besides human-labor surveys that are subjective, time-consuming, and expensive, autonomous image and video analysis is a faster, more efficient, and non-destructive way. This thesis focuses on crack detection from videos, crack segmentation from images, and building assessment from street view images. For crack detection from videos, three approaches are proposed based on local binary pattern (LBP) and support vector machine (SVM), deep convolution neural network (DCNN), and fully-connected network (FCN). A parametric Naïve Bayes data fusion scheme is introduced that registers video frames in a spatiotemporal coordinate system and fuses information based on Bayesian probability to increase detection precision. For crack segmentation from images, the rotation-invariant property of crack is utilized to enhance the segmentation accuracy. The architectures of several approximately rotation-invariant DCNNs are discussed and compared using several crack datasets. For building assessment from street view images, a framework of multiple DCNNs is proposed to detect buildings and predict their attributes that are crucial for flood risk estimation, including founding heights, foundation types (pier, slab, mobile home, or others), building types (commercial, residential, or mobile home), and building stories. A feature fusion scheme is proposed that combines image feature with meta information to improve the predictions, and a task relation encoding network (TREncNet) is introduced that encodes task relations as network connections to enhance multi-task learning.