to improve construction quality. The focus is not on replac- ing construction inspectors but on enhancing their ability to conduct global assessments quickly and in a thorough fashion by combining low-cost technologies available in the market with an artificial intelligence framework. Specifically, this paper emphasizes using new low-cost technology to obtain the rebar size in reinforced concrete construction. This study in con- junction with work of other researchers in rebar spacing and cover estimation can automatize the rebar geometric inspec- tion. It can also be used as a general concept for inspection automation, where the collection of the data and its analysis can eventually inform the projected strength of the structure in real time. This also facilitates moving towards e-construc- tion, where quality construction is achieved using a paperless format while also creating a permanent record. Technology for Structure Inspection Traditional infrastructure inspections are performed by qual- ified inspectors who physically go to the field and perform inspections (Newman and Jain 1995). In the industrial, com- mercial, and civil sectors of the construction industry, the application of automation technologies has attracted attention in recent years. These technologies, such as uncrewed aerial vehicles (UAVs) or drones (Kuo et al. 2016), LiDAR (Wood and Mohammadi 2015), augmented reality (AR) (Wang et al. 2019), and RGBD cameras (Wójcik and Zarski . 2021) are being used for collecting 3D geometric information for inspection, renovation, and retrofit projects. Researchers use UAVs to reconstruct a 3D image of the structure using a camera and remote sensor (Chen et al. 2019 Perry et al. 2020 Ayele et al. 2020). UAVs can be used in areas that are hard to access, but they have a limitation of payload equipment, vibration, environmental condition, and cost. LiDAR is another instrument that researchers consider an inspection tool, and it has been used to automatically evaluate the rebar spacing before pouring concrete (Yuan et al. 2021a). In another study, they proposed a platform for mounting the LiDAR on the UAV to detect discontinuities of the structure (Nasrollahi et al. 2018). Despite the high accuracy of collected data from the LiDAR, it requires an expert to be familiar with the technology. Also, the setup time in the field is, in general, a concern for construction managers and structural engineers who need a quick quality control/quality assurance method to inform them in a short time on the rebar quality. AR is another new technology that has recently attracted researchers’ atten- tion. There are studies in which the authors used the AR-BIM platform for inspection of the bridge and locating defects (John Samuel et al. 2022). Also, the application of AR in infrastructure inspection and its capability have been examined in the past (Mascareñas et al. 2021). The rapid development of low-cost RGBD sensors has attracted the attention of researchers in various scientific and technological fields due to the advan- tages they offer. The RGBD sensor includes a color sensor and a depth sensor that captures color information and depth data by measuring the time between emitted and reflected light bouncing back to the visor (time of flight). An RGBD camera was used in the literature to detect the position of a target and its orientation on a bridge that was identified as damaged (Ivanovic et al. 2021). Additionally, other important factors such as spacing and cover of the rebar need to be controlled before pouring the concrete. A group of researchers used an RGBD camera to obtain the rebar spacing and cover in a rebar cage automatically to demonstrate the capability of a low-cost approach to assist in quantifying the quality of rebar place- ment. This method was tested on a rebar cage to calculate the spacing and the result was validated with tape measurement (Yuan et al. 2021b). In addition to the capability of measuring rebar spacing and clearance, it would be of value to explore if the proposed methodology would be able to measure the rebar diameter automatically. Each of these instruments has advantages and disad- vantages that make the engineer and owner use them based on the project’s specifications. However, there are important factors to consider for all inspectors as they can enhance the speed, cost, and quality of the inspection. Handheld, portable, low-cost instruments are among those capabilities that are worth considering. Inspection Industry Standards For more than a century, reinforced concrete has dominated the global construction industry, even though RC structures are frequently subject to a variety of deterioration and damage owing to various exposure factors. As a result, the require- ments of the building standards necessitate somewhat frequent inspections in many countries (Masoumi et al. 2013). The emphasis of the proposed method is on the inspection of rebar in RC structures, such as buildings or bridges. Rebar placement has a significant impact on the quality of RC structures (Qi et al. 2014). Rebar defects in concrete structures can induce struc- tural collapse, uneven strength capacity, and serious concrete degradation (Cusson 2009). Incorrect rebar placement or insufficient concrete cover might decrease structural strength. One of the main issues limiting the durability of old RC struc- tures is the rebar corrosion brought on when the concrete cover is not sufficient (Wilkie and Dyer 2022). Due to this, the concrete structure becomes less durable and is more suscepti- ble to chemical attacks (Chemrouk 2015). RC structures benefit from accurate rebar arrangements for the duration of their lives (Słowik 2019 Concrete Construction 2005). Rebar construction errors can be reduced to a minimum to avoid costly repairs, task deferrals, flaws in structural strength, and even save struc- tures from collapsing. Current Inspections Limitations Figure 1 shows the labor during rebar placement at a bridge deck construction project. Projects of this size can take several days to weeks for rebar placement. In tight construction sched- ules recognizing issues with placement as the project progresses is important in assuring that issues are addressed as soon as possible. Early detection of errors can accelerate construction. J A N U A R Y 2 0 2 3 • M AT E R I A L S E V A L U AT I O N 47 2301 ME Jan New.indd 47 12/20/22 8:15 AM

It is important to acquire geometric data for infrastruc- ture management (Concrete Construction 2005). The present method for collecting geometric structure data for inspection of the structure relies on manual data collection methods, which require time and could be prone to error (Rens et al. 1997). Inspectors that are close to the hard-to-reach parts of the bridge (Kuang et al. 2009) may face safety issues, making the data collection process cumbersome and time-consuming (Sanford et al. 1999). Furthermore, the instrumentation can take measurements only in discrete and sparsely populated areas. In addition, most of the collected data is recorded in the form of paper-based or written files without any support, so the process of interpreting and retrieving geometric infor- mation requires a large amount of data transmission work to be done manually by a human (Estes and Frangopol 2003). Therefore, it would be advantageous to record the final location of reinforcement as a future document to be acces- sible if needed during the service life. This would become a permanent record of the field. When a forensic investigation is needed, a potentially effective tool for measuring concrete cover after construction could be ground penetrating radar (GPR), which uses penetrating wave. Researchers and com- panies use GPR to measure the actual cover, spacing, and bar diameter (Tosti and Ferrante 2020). Not only are GPR and other means expensive and only justified in specific cases, but also it is not always possible to use them due to traffic or other functionality restraints (De Souza et al. 2004). The cost of not knowing the actual quality of the structure will translate into negative consequences both in cost and safety (BrÜhwiler 2009). Hence, an automated way of quantifying reinforcement in structures such as bridge decks using innovative technologies coupled with artificial intelligence (advanced processing and decision-making concepts) is agreed to be useful for construc- tion quality (Akinci et al. 2006). The proposed method can inform the inspector of the reinforcement during construction and become a permanent record of the quality that can be accessed in the future. Methodology This section describes the technical capabilities of the Azure Kinect camera in relation to data collection, as well as examples of RGBD applications used in the past. Subsequently, the capability of using RGBD cameras for creating 3D point cloud data for rebar inspection is explained. RGBD Camera The RGBD camera is a color-depth camera that has the poten- tial of reconstructing the 3D shape of an object. There are different types of RGBD cameras that can be used for this study. In 2020, Microsoft launched the third generation of Kinect, called Azure Kinect, with time of flight technology. This sensor is offered at a relatively low price (currently less than US$1000) but offers practical performance and depth accuracy. 3D scanning and measurement capabilities bring low-cost Kinect sensors to a wider range of applications such as gesture recognition (Ren et al. 2013 Patsadu et al. 2012 Biswas and Basu 2011), AR (Ren et al. 2013 Ma et al. 2013 Vera et al. 2011), robot navigation (Correa et al. 2012), and the construction of the building (El-laithy et al. 2012). Figure 2 shows the RGBD camera that is used in this study. Figure 2 shows Microsoft’s newest sensor technology that is incorporated into the Kinect hardware, which is a single USB- connected device. The camera dimensions and weight are 103 × 39 × 126 mm and 440 g, respectively. This RGBD camera ME | RGBDCAMERAS Figure 1. Rebar mat inspection on a big bridge deck construction site. Figure 2. Azure Kinect camera: (a) front view (b) lateral view. 103 mm 39 mm 126 mm 48 M AT E R I A L S E V A L U AT I O N • J A N U A R Y 2 0 2 3 2301 ME Jan New.indd 48 12/20/22 8:15 AM COURTESY: XINXING YUAN