Abstract
Background: The City of Johannesburg (CoJ) is grappling with deteriorating road infrastructure marked by potholes, surface cracks, faded markings and blocked drainage. Traditional visual inspections are reactive, subjective and labour-intensive.
Aim: This study explores the use of machine vision technologies to enhance road condition assessment in CoJ. The aim is also to analyse regional road damage patterns; evaluate the effectiveness of machine vision in detecting and classifying defects; develop and validate an automated framework for defect detection and maintenance planning and recommend adoption of advanced technologies to improve inspection accuracy.
Setting: The study was conducted within CoJ Metropolitan Municipality, South Africa.
Methods: A mixed methods approach was used using Chi-square tests, analysis of variation (ANOVA), correlation and structural equation modelling (SEM), complemented by qualitative insights.
Results: Significant regional variations were observed. Potholes were prevalent in the Regions A, E and G, while surface cracks dominated in the Regions B and C. A machine vision framework was developed to automate road damage detection.
Conclusion: The study concludes that machine vision improves the accuracy and efficiency of road condition assessments.
Contribution: The study delivers a scalable framework for municipal road management and informs smart city policy and investment.
Keywords: machine vision; road condition assessment; artificial intelligence; urban infrastructure; Johannesburg, pavement management; predictive maintenance.
Introduction
In the City of Johannesburg (CoJ) Metropolitan Municipality, road condition issues such as potholes, cracks, fading road markings, uneven surfaces and inadequate drainage are critical service delivery challenges aggravated by ageing infrastructure and heavy traffic (Duvenage 2025). The increasing complexity of urban infrastructure maintenance, particularly in large metropolitan municipalities, demands more efficient and intelligent methods of assessing road conditions. Traditional manual inspections are often resource-intensive, reactive and limited in scope, leading to suboptimal maintenance planning and delayed service delivery. Yang et al. (2024) state that the current road condition assessment techniques heavily rely on manual road condition assessment methods and reactive maintenance, which suffer from a lack of comprehensive coverage and are both costly and resource-intensive. To address these challenges, machine vision solutions such as automated real-time monitoring systems, advanced image processing algorithms, predictive maintenance tools, supported by cloud-based data management, could significantly enhance the efficiency and accuracy of road condition assessments (Yang et al. 2024). These technologies would enable a more proactive approach to road maintenance, improving safety and optimising resource use across the CoJ’s extensive road network (Hancock 2022). Salcedo, Jaber and Requena Carrión (2022) state that by integrating machine vision, the analysis is enhanced with precise, real-time data on road conditions, while Moyo et al. (2022) add that this will facilitate a deeper understanding of the spatial and demographic correlations with road maintenance issues across all seven regions of the CoJ. In their integrated development plan, CoJ (2025) indicates that the goal is to utilise insights to propose data-driven strategies tailored to enhance urban infrastructure management, ensuring equitable and effective road maintenance across the Metropolitan Municipality. This approach not only aims to improve current conditions but also to foster a more sustainable urban environment through strategic planning and the innovative use of technology in maintenance practices (CoJ 2025). This study explores the complex issue of road maintenance in the CoJ Metropolitan Municipality, a pivotal urban hub in South Africa (Rose 2024), grappling with infrastructure challenges that affect its varied demographic segments.
The significance of this study lies in its response to pressing urban challenges faced by the CoJ Metropolitan Municipality, deteriorating road infrastructure that directly impacts service delivery, public safety and the quality of life for residents across different regions of the city. Overall, this study offers a novel and contextually relevant contribution by integrating technical innovation, policy application and equity considerations within the domain of urban infrastructure management, thereby distinguishing itself from the prevailing body of literature.
Literature review
Overview of road issues and damages in South Africa
South Africa has a road network spanning around 750 000 km, making it the longest among all African nations and ranking it as the seventh longest globally. The South African National Roads Agency Limited (SANRAL) is responsible for the management of the country’s national road infrastructure, which encompasses a vast network spanning approximately 22 197 km of paved highways (Sanral 2024). Provincial governments are responsible for over 220 000 km, while the municipal network is approximately 275 000 km. The CoJ Metropolitan Municipality has about 13 599 km of roads, with over 91% which are surfaced and 9% unsurfaced (Hancock 2022). According to the 2022 CoJ Metropolitan Municipality annual report, 45% of the road network is in excellent condition, 23% is poor and 32% is extremely poor (City of Johannesburg 2022). Potholes, structural cracks, faulty streetlights, clogged storm water drains, neglected walkways and traffic islands are signs of poor road conditions. As a result of storm-water drain obstruction, water flows along heavily used streets, making them unusable by vehicles and forming new potholes and expanding existing ones (CSIR 2023). The CoJ Metropolitan Municipality has fixed up to 100 000 potholes in 9 months in 2023 and road users report at least 1000 potholes every day (Peters-Scheepers 2023).
Overview of road condition assessment methods used in South Africa
The constant and essential process of maintaining road infrastructure is crucial to retaining the quality, safety and operation of road surfaces. Afridi, Erlingsson and Sjögren (2023) suggest that to ensure road assets are preserved and properly maintained, government departments should carry out regular road condition assessments. Chen et al. (2022) further recommend adopting proactive strategies to mitigate the deterioration of road infrastructure caused by various factors such as ageing and usage over time. Road condition assessments are used to perform a comprehensive evaluation of the road surface to detect the presence of cracks, potholes, rutting and signs of degradation (Pavel, Tan & Abdullah 2022). The assessment of road conditions using traditional methodologies is carried out by the utilisation of manual tools, accurate measuring devices and visual inspections (Pavel et al. 2022). According to Pavel et al. (2022), the process of visual inspection requires experts to evaluate the state of road infrastructure components by using data obtained via visual inspections. This process is fundamentally subjective, as it mostly relies on the evaluative opinions and expertise of the inspectors (Ren et al. 2022). Various inspectors with diverse backgrounds and perspectives may potentially perceive occurrences in different ways, thus resulting in disparities or variations in their assessments (Jia et al. 2023). According to Ahmad et al. (2024), relying just on visual inspections may be inadequate in detecting structural or subsurface issues that might potentially compromise the overall integrity of the highway. Visual inspections are occasionally limited to superficial analysis and may not sufficiently detect concealed flaws (Afridi et al. 2023). The visual inspection evaluation of road conditions encompasses several methodologies such as the pavement condition index (PCI), the international roughness index (IRI), functional categorisation and network-level assessment (Ren et al. 2022).
Pavement condition index
The PCI is a subjective-type method for visually evaluating pavements, which considers the extent and severity of the defects found in a particular inspection site (Ren et al. 2022). According to Huang et al. (2024), the PCI is a method used to evaluate the state of road pavements based on their classification and the extent of damage. Jia et al. (2023) define PCI as a technique used for assessing the condition of road pavements, taking into consideration their classification and the level of damage sustained. Hancock (2022) states that PCI serves as a useful tool for maintenance purposes. The PCI is a systematic approach for visually assessing the condition of pavements. The PCI method classifies sections using PCI values in the ranges of 0–100, which covers seven different road surface conditions, on the maintenance priority (good, satisfactory and fair) and the rehabilitation priority levels (poor, very poor, serious and failed). The PCI allows for the assessment of the state of the pavement by considering the defects present on its surface (Li & Hsu 2022). Li and Hsu (2022) state that the PCI offers a subjective approach to evaluating the repair process of a specific pavement and determining the order of importance for maintenance jobs.
International roughness index
The IRI is a metric used to quantify the roughness of roads (Lin et al. 2022). The IRI is a measure that quantifies the cumulative elevation difference of a single wheel path on a test road segment divided by the length of the test road (Javaid et al. 2022). The IRI represents the range of road height abnormalities per metre or kilometres in a determined road section (Pinatt et al. 2020). Lin et al. (2022) state that IRI serves as a reference average slope correction for evaluating road roughness. The IRI features time stability, serves as an objective indicator of pavement conditions and offers freedom from human-factor interference (Deni 2020). Lin et al. (2022) outline how the IRI measures the unevenness of a road, in that it is reported in metres per kilometre (m/km). Lin et al. (2022) further state that the measurement quantifies the total vertical movement encountered by a vehicle’s suspension system when traversing a road surface. A higher IRI value indicates a rougher road surface, whereas a lower IRI value indicates a smoother road surface (Lin et al. 2022). The major objective of this system is to evaluate the roughness of the road surface and aid in the scheduling and implementation of regular preventative maintenance (Pinatt et al. 2020).
Road network-level assessment
Road network-level evaluation is a methodology that aims to review the comprehensive state and functionality of the entire road network, as opposed to isolated road segments (Wintruff & Fernandes 2023). Pavel et al. (2022) clearly articulate that the main objectives of evaluations are to assess the current conditions; identify maintenance; preservation and rehabilitation projects; prioritise projects and allocate budgets and determine funding needs. Moolla and Tetley (2021) state that network-level assessment offers a comprehensive evaluation of the status of the road network; however, Pinatt et al. (2020) raised a point that network-level assessment does not give the particular and detailed information required to effectively address maintenance and repair requirements at the individual segment level. Chen et al. (2022) state that a comprehensive evaluation conducted at the network level may not effectively identify specific concerns that are localised to individual road segments that require urgent action, such as significant road surface deterioration, safety risks or drainage difficulties. Furthermore, Wintruff and Fernandes (2023) state that the gathering of performance data on a road network over a period serves as a helpful instrument for monitoring the quality of road surfaces and as a means for constructing performance models that may be utilised to forecast future circumstances.
Pavel et al (2022) highlight the fact that physical visual assessment requires experienced and skilled human resources, is dangerous, stressful and time-consuming with production of ± 60 km –80 km per day in a rural environment and up to 20 km per day for urban roads. Modern road network assessments have evolved from manual human-based visual survey, which are slow, subjective and labour-intensive to technology-driven methods that provide fast, objective and repeatable results. Key technologies include the Traffic Speed Deflectometer (TSD), which measures pavement deflection and structural capacity continuously at normal traffic speeds without lane closures, the Falling Weight Deflectometer (FWD), which applies a stationary impulse load to measure pavement strength at discrete points and Ground Penetrating Radar (GPR) for layer thickness and subsurface conditions. While human assessments remain useful for targeted verification, technology-based methods are now preferred for network-level surveys because they reduce long-term costs, allow faster coverage and enable integration with Geographic information system (GIS)-based pavement management systems. For large road networks, TSD and mobile imaging systems are more cost-effective because of their speed and continuous data capture, while FWD is more suited to project-level spot investigations, manual surveys, despite low initial equipment cost, incur higher labour costs and are less scalable. Wintruff and Fernandes (2023) further highlight that the activities that may be carried out are frequently determined by the available resources.
Weakness of road condition assessment methods used in South Africa
Although the PCI is a valuable tool for evaluating pavement conditions, Hancock (2022) remarks it does have several drawbacks. Hancock (2022) states that PCI is not suitable as an engineering tool for local governments when making choices regarding pavement management because of its limitations. More precisely, when a PCI is created using data from condition surveys, Lendra et al. (2023) outline that a significant amount of crucial engineering information is omitted, specifically data related to cracking. One significant drawback of PCI is that roadway segments might exhibit comparable or identical PCI values while having distinct types of distress (Lendra et al. 2023). Lin et al. (2022) are of the view that performing PCI surveys necessitates a substantial amount of time and effort, particularly for extensive networks.
While IRI is a widely used index for indicating road surface roughness, there are some limitations (Lin et al. 2022). The measurement of IRI values has traditionally relied on a manual method, which presents several drawbacks such as time-consuming procedures, traffic disruption and relatively high costs (Wintruff & Fernandes 2023). Pinatt et al. (2020) present a view that manual methods may be cost-effective but can be time-consuming and subject to user bias. Moolla and Tetley (2021) state that IRI tools provide objective deflection data, but require specialised equipment and may cause disruptions in traffic flow. Moolla and Tetley (2021) also state that during the IRI assessment, inaccuracy in the number of vehicles’ wheels can affect the accuracy of IRI values. The relationship between IRI and other measures of roughness and riding quality is not always clear (Lin et al. 2022). Lin et al. (2022) highlight that the inspection of surface roughness may not detect underlying structural problems. While passenger comfort is important, the methodology lacks a comprehensive assessment of other aspects related to road conditions, such as road surface degradation, safety issues and structural soundness (Afridi et al. 2023). Ahmed, Kays & Sadri (2023) say that IRI offers a level of stability that guarantees that trends and variations in pavement roughness may be precisely monitored and compared on an annual basis. As a result of its inherent stability, Ahmed et al. (2023) say that the IRI is a dependable tool for forecasting future pavement conditions. Jia et al. (2023) add that it facilitates the planning of maintenance and rehabilitation operations. International roughness index is a universally acknowledged and globally utilised standardised approach for quantifying road roughness (Wintruff & Fernandes 2023).
According to Starke and Geiger (2022), network-level assessment involves a certain degree of subjectivity when estimating distress amounts. Ren et al. (2022) say that if the inspectors are not well qualified, there is a higher chance of having more variability in the results compared to if distress quantities had been directly measured. At the network level, the significance lies in matters pertaining to the amount and quality of data (Afridi et al. 2023). The measurement of data quantity, in terms of both the type of data being measured and the amount being measured, has consequences for both time and cost (Afridi et al. 2023). Typically, the cost of data collecting increases as the volume or level of detail of the acquired data increases (Pereira & Vieira 2022). Simultaneously, Wintruff and Fernandes (2023) are of the view that the availability of additional or more comprehensive data for analysis would lead to improved decision-making.
Although road network-level assessment offers numerous benefits, it also presents certain drawbacks and challenges (Afridi et al. 2023). Chen et al. (2022) state that performing thorough evaluations of extensive road networks can incur high costs; Afridi et al. (2023) add that it necessitates the use of substantial financial resources for the procurement of advanced technology, equipment and human resources. Jia et al. (2023) add that processing and evaluating the vast quantities of data gathered during assessments can be intricate and demanding in terms of resources. According to Lin et al. (2022), network-level assessment requires advanced data management systems and experienced workers. Wintruff and Fernandes (2023) state that maintaining the precision and uniformity of data throughout a vast network might pose difficulties. Erroneous or contradictory data can result in incorrect decisions and unsuccessful solutions (Lin et al. 2022). Pinatt et al. (2020) found that the evaluation procedure, particularly when it entails physical inspections, might result in temporary interruptions to traffic flow and local populations. The efficiency of road network-level assessments can be hindered by the scarcity of competent persons and innovative technology in certain places (Deni 2020). Lin et al. (2022) add that the process of incorporating new evaluation data into current transport management systems and databases can be difficult and may necessitate extra resources and time.
Machine vision
Machine vision is an automated system that uses optical devices and non-contact sensors to capture and analyse images of real objects (Schattler, Wolters & Zimmerman 2020). Machine vision can perform a diverse range of tasks that human vision is unable to handle (Moolla & Tetley 2021). Schattler et al. (2020) outline that performing capturing of images is achieved through the utilisation of advanced vision sensors, cleverly designed optical transmission systems and powerful image processing algorithms (Figure 1). In the domain of machine vision, images function as the conveyors of information. Image processing and evaluation are crucial technologies used in vision detection systems to automatically comprehend received images (Moolla & Tetley 2021). The use of computer vision technology allows for the creation of machine vision, which is used to detect surface objects.
 |
FIGURE 1: Architecture Of Machine Vision System. |
|
Within industrial applications, there are three distinct defect detection tasks that rely on machine vision: classification, localisation and segmentation (Kitsios & Kamariotou 2021). Kitsios and Kamariotou (2021) argue that in most cases, defect classification is utilised to determine whether a particular fault is present in a picture (Cazzato et al. 2020). Cazzato et al. (2020) state that information that specifies the features of the target is extracted from the image pixels through the process of feature extraction. After that, the differences between the various targets are mapped to a feature space with a reduced dimension. For determining the categories of defects that are present in an image, it is critical that the features that are picked to describe the image distinguish between different types of images (Kitsios & Kamariotou 2021). The fundamental objective of defect classification is to train the classifier in accordance with the extracted feature set and then to ensure that it correctly identifies the classification of surface defects for each surface defect using either supervised or unsupervised pattern recognition techniques. Defect localisation involves the use of imaging techniques and algorithms to accurately locate and detect defects in objects. In the specific context of a picture, fault localisation necessitates accurately identifying the defect’s position and categorising it accordingly (Cazzato et al. 2020). Defect localisation is the process of identifying all the specific locations within the product where defects are present. Furthermore, significant advancements have been achieved in defect localisation by utilising Convolutional Neural Networks (CNN)-based image classification techniques and deep learning-based object detection technology (Abdollahi et al. 2020). Image segmentation is a method that divides images into different sections that are distinct from one another for the purpose of identifying objects of interest with the main objective of predicting categories that each pixel in an image belongs to (Cazzato et al. 2020). Regional-based, edge-based and specialised theory-based approaches are used for the process of segmentation.
Machine vision in road asset management
Machine vision has greatly enhanced road inspection procedures by offering more precise, efficient and cost-efficient solutions in contrast to conventional approaches (Abdollahi et al. 2020). Akinosho et al. (2020) further state that machine vision systems have the capability to identify and categorise cracks, potholes and other imperfections on road surfaces. Khan et al. (2020) indicate that advanced cameras with high quality capture photos or videos, while sophisticated machine learning algorithms analyse them to accurately identify any faults. Huang et al. (2023) add that automated devices evaluate the depth and intensity of grooves in the road surface, which are vital for evaluating road deterioration and strategising repair. According to Pinatt et al. (2020), machine vision can assess the pavement surface’s texture, aiding in the determination of skid resistance and overall road safety. Kim et al. (2023) provide input that visual inspection systems can identify indications of material degradation, such as alterations or modifications in surface composition, which serve as indicators for the requirement of maintenance or resurfacing. Moolla and Tetley (2021) further support the use of machine vision because the systems can evaluate the state and perceptibility of road markings and signage, verifying that they comply with safety regulations and are easily noticeable to drivers. Kim et al. (2022), Huang et al. (2023) and Shah et al. (2023) have indicated that automated devices evaluate the retroreflectivity of road markers and signs to verify their adequacy in reflecting light for night-time vision. State-of-the-art machine vision systems generate three-dimensional profiles of road surfaces, enabling the detection of unevenness, bumps and other anomalies (Lemonakis, Kopelias & Karlaftis 2023; Salcedo et al. 2022). By conducting a comparative analysis of three-dimensional (3D) profiles over a period, Schattler et al. (2020) are of the view that it is possible to track the rate of deterioration, which Marianingsih et al. (2022) agree can be helpful in facilitating proactive maintenance.
There are a few advantages to employing machine vision. Burghardt et al. (2021) mention that machine vision systems offer a high level of precision when it comes to detecting and quantifying road problems. Akinosho et al. (2020) agree that machine vision assessment results in more accurate evaluations when compared to manual inspections. El Hakea and Fakhr (2023) reviewed 190 studies on computer vision applications for pavement distress and condition assessment and found that research in this field has grown rapidly, with deep learning approach replacing traditional image processing techniques. El Hakea and Fakhr (2023), Shah et al. (2023) and Wang et al. (2022) mention that automated inspection systems can cover extensive road sections in a short amount of time. Shah et al. (2023) state that machine vision helps to reduce overall maintenance costs by lowering the amount of manual labour that is required and by increasing the pace at which inspections are performed. Machine vision reduces inspection time by approximately 70% – 90% compared to manual surveys because high-speed cameras, LiDAR (Light Detection and Ranging) and automated image processing can collect data while travelling at normal traffic speeds, covering 200 km – 500 km per day, whereas manual visual surveys typically achieve only 20 km – 50 km per day at a network level and 5 km – 15 km per day for detailed assessments. Schattler et al. (2020) comprehend that automated inspections eliminate the need for human inspectors to be present on the road, which improves safety by reducing the amount of time spent in danger from potential traffic hazards.
Research methods and design
This study used a sequential mixed-methods research design, which was particularly appropriate because of the multifaceted approach required to address the complex issues of road condition assessment within the CoJ. The research began with a quantitative analysis of the road conditions across various regions within the CoJ. This initial phase involved collecting and numerically analysing data to identify prevalent issues and patterns in road degradation (Figure 2). Following the quantitative analysis, the study transitioned to a qualitative phase where in-depth inquiries were conducted.
The incorporation of qualitative techniques was deemed essential to provide deeper, contextual insights into the systemic, operational and experiential factors underpinning the quantitative findings, thereby enabling a better understanding of the patterns of road degradation identified in the initial phase. The qualitative component followed a rigorous and structured process: key stakeholders, including municipal engineers, maintenance personnel and road users, were purposively selected based on the outcomes of the quantitative analysis; a semi-structured interview guide was meticulously developed to facilitate focused yet flexible data collection; ethical clearance was obtained and informed consent was secured from all participants. Data were collected through in-depth interviews, focus group discussions and direct field observations, and subsequently transcribed verbatim to ensure data integrity. Thematic analysis was conducted using a systematic coding process to identify recurring themes and patterns. To enhance the validity and reliability of the findings, triangulation was employed by cross-verifying interview data with field observations and documentary evidence such as maintenance records.
The target participants for the questionnaire were primarily municipal road engineers and maintenance staff who are directly involved in road assessment and repairs within the CoJ.
Slovin’s formula was used to calculate the minimum sample size needed to estimate a statistic based on an acceptable margin of error. Slovin’s formula is calculated as: n = N/ (1 + Ne2), where: n = sample size for participants to achieve a 95% confidence level with a 5% margin of error; N = population size of the study; and e = acceptable margin of error. Therefore, the sample size n = 1771/(1 + 1771 × 0.052) = 326. The initial proportional framework indicated that a statistically representative sample for the entire population would require approximately 326 participants. However, given the exploratory nature of this mixed-methods study, which combines quantitative surveys with qualitative interviews and focus groups, it was determined that a reduced, yet purposive sample of 100 participants would be appropriate. The selection was informed by established guidelines for qualitative research (Fusch & Ness 2015; Hennink & Kaiser 2022), where the emphasis is on data saturation rather than statistical representativeness. The distribution of the questionnaire link was conducted via email and WhatsApp, leveraging the internal communication channels of the CoJ road department and supporting agencies. This method was selected for its efficiency and ability to reach a targeted group of participants within a reasonable timeframe. Thematic analysis was used to interpret qualitative responses. Descriptive statistics were used to have an overview of the demographic characteristics of the respondents by performing frequencies on categorical demographic variables such as gender and ethnicity. Inferential statistics were performed using correlation analysis and structural equation modelling (SEM). Hidayat and Wulandari (2022) affirm that SEM is a statistical methodology employed for the examination and representation of intricate associations between observable and latent variables. Structural equation modelling integrates components of factor analysis and multiple regression analysis to investigate indirect associations, offering valuable insights into intricate causal interactions among numerous variables. A threshold of 0.05 was utilised as the significance level for establishing statistical significance. A multiple regression model was used to examine the relationship between the dependent variable (road conditions assessment) and independent variables (Machine learning, Machine vision, Expert System and Natural Language Processing (ESNLP) and Robots. The Regression Sum of Squares was used to measure how much of the variability in the dependent variable is explained by the regression model. Finally, the importance–performance was utilised to determine the most influential independent variable on road condition assessment quality. Chi-square (χ2) Test was conducted.
Ethical considerations
Ethical approval to conduct this study was obtained from the Tshwane University of Technology Research Ethics Committee (No. REC2024=04=039(MS)). Strict ethical considerations were adhered to throughout all stages of the research. Ethical clearance was obtained from the CoJ before data collection commenced. All participants were informed of the purpose of the study, their voluntary participation and their right to withdraw at any time without consequence. Informed consent was obtained prior to administering questionnaires, interviews and focus groups. To protect confidentiality, participants’ identities and responses were anonymised, and all collected data were stored securely and used solely for academic purposes.
Results
Demographic overview
Table 1 represents the demographics of research participants within the seven regions of the CoJ Metropolitan Municipality.
The demographic profile presented in Table 1 indicates distinct distributions within gender and racial categories for a specified population. Gender distribution is characterised by a predominant proportion of males, constituting 64.2%, while females represent 34.6% and an additional 1.2% of individuals identify as other, suggesting recognition of diverse gender identities. In terms of racial composition, most of the population is black, accounting for 87.7%, followed by coloured individuals at 8.6% and Indian/Asian at 3.7%. These data highlight a demographic landscape where black individuals significantly outnumber other racial groups, with meaningful representation from other racial backgrounds.
The occupational profile in Table 2 presented a distribution across different organisational levels. Executive managers, comprising chief executive officer (CEOs), chief financial officers (CFOs), chief operating officers (COOs) and heads of departments (HODs), represent 0.85% of the sample, reflecting a small, but strategically significant segment of leadership. Senior managers and directors account for 2.48%, while managers and deputy directors, predominantly registered professional engineers, constitute 17.56% of the participants. The largest proportion of the sample consists of skilled employees, including administrators, artisans and technicians, who make up 79.11% of the population.
| TABLE 2: Distribution of purposeful sampling of the study. |
Types of road damage experienced in different regions of the City of Johannesburg
Figure 3 illustrates the geographic demarcation of the seven administrative regions of the CoJ, labelled A through G. These regions represent spatial zones within the metropolitan boundary, each encompassing a range of suburbs, townships and industrial areas with differing socio-economic characteristics.
 |
FIGURE 3: Map of Johannesburg showing the seven regions of the city. |
|
Notably, potholes predominate in Regions A, E and G, constituting 80.0%, 60.0% and 75.0% of the reported damages, respectively, which suggests significant road infrastructure distress. Surface cracks are predominantly observed in Regions B and C, accounting for 83.3% and 54.5% of damages, respectively, indicative of potential underlying issues such as material failure or subgrade instability. Other categories of damage, including worn lane markings and water flooding, appear less frequently yet are evident in Regions C through E, highlighting concerns related to visibility and drainage efficiency. In addition, uneven surfaces are specifically prevalent in Region F, representing 13.3% of the reported damages, potentially because of inadequate construction practices or the impact of heavy vehicular traffic. These observations underscore the necessity for region-specific road maintenance and rehabilitation strategies that directly address the predominant forms of damage in each area, thereby enhancing the overall safety and durability of road infrastructure.
Figure 4 represents common types of road damage in seven regions of CoJ indicating the distribution of various types of road damage across regions labelled A through G, indicating substantial regional discrepancies.
 |
FIGURE 4: Types of road damages in the City of Johannesburg Metropolitan Municipality. |
|
Statistical significance of test results for the road damage experienced in different regions of the City of Johannesburg
Statistical analyses were used to evaluate the validity, reliability and overall robustness of the study’s findings. The Chi-square analysis validated that the observed associations among variables were statistically significant and improbable to have arisen by chance. The Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s Test of Sphericity were used to confirm the sample adequacy and inter-variable correlations that justified advanced statistical analysis. The analysis of variance (ANOVA) inside the regression model demonstrated that the implementation of artificial intelligence (AI) technologies considerably and consistently influenced the quality of road condition evaluations. Table 3 represents the results of a Chi-square Test that was conducted to analyse the frequency distribution of road damage across different regions. The test aims to determine if there are statistically significant differences in the types of road damage reported in the various regions of the CoJ.
The analytical findings from the Chi-square Test reveal significant variations in the types of road damage across different regions, with a p-value of 0.004. This statistical significance highlights distinct regional challenges, suggesting that environmental, administrative or material quality differences could be influencing the prevalence and types of road damage observed. Table 4 represents KMO test for suitability factor analysis.
| TABLE 4: Kaiser–Meyer–Olkin and Bartlett’s tests. |
The application of the KMO and Bartlett’s Tests to evaluate the dataset’s suitability for factor analysis yielded a KMO value of 0.627. This result not only confirms the adequacy and acceptability of the sample size but also indicates that the data are appropriate for detecting underlying patterns between variables. These analyses collectively provide a robust statistical framework to further explore and understand the dynamics of road damage, potentially guiding targeted interventions and resource allocation for infrastructure maintenance and improvement.
Table 5 (ANOVA) evaluates the effectiveness of a regression model in explaining variations in road condition assessments, using predictors such as Robots, ESNLP, machine learning and machine vision.
Table 5 shows a Regression Sum of Squares at 37.504, indicating the variability explained by the model and a Residual Sum of Squares at 21.466, representing unexplained variability. The total variation in the dataset is 58.970. With five predictors, the model has an F-value of 107.971, suggesting a highly significant fit (p < 0.001), confirming that the predictors substantially and significantly contribute to predicting road conditions assessment, with the mean square values (7.501 for regression and 0.069 for residual) further supporting the model’s explanatory power.
The multiple regression analysis was carried out using Statistical Package for the Social Sciences (SPSS) Version 24. According to the final model, machine vision, ESNLP, and robots with beta coefficients of 0.193, 0.165 and 0.269 for X1 (machine vision), X2 (ESNLP) and X3 (robots), respectively, are critical factors for road condition assessment.
The developed mathematical equation of the model is as follows (Equation 1):
Where:
RCA = Road condition assessment;
X1 = Machine Vision (MV) with β = 0.193 and p-value < 0.05;
X2 = Expert System and Natural Language Processing (ESNLP) with β = 0.165 and p-value < 0.05;
X3 = Robots (R) with β = 0.269 and p-value < 0.05.
Machine vision framework
Figure 5 represents a machine vision framework for road conditions assessment.
The Machine Vision-Based Road Condition Assessment Framework automates road monitoring using AI and GIS mapping for efficient maintenance planning. It begins with data collection from vehicle-mounted cameras, drones, satellite imagery and citizen reports. The images undergo preprocessing and augmentation, including normalisation, region of interest extraction and enhancement for better model accuracy. A machine learning model (using CNNs, You Only Look Once [YOLO] or Mask R–CNN) detects road defects such as potholes and cracks through object detection and semantic segmentation. The identified defects are then geo-tagged and mapped using a GIS system, creating real-time dashboards and heatmaps. The decision support system analyses these data to prioritise road repairs, estimate costs and predict future maintenance needs. Finally, a Road Condition Assessment report is generated, providing an overall road condition score, identifying high-priority areas, recommending repair actions and assisting in budget planning for Johannesburg Municipality. This framework ensures a data-driven, proactive approach to road maintenance, reducing manual inspections and improving resource allocation.
Qualitative analysis
The qualitative phase highlighted critical deficiencies in the current system, including reliance on manual and subjective assessment methods, outdated tools and techniques, inconsistent record keeping, slow response times and insufficient integration of data into planning processes (Table 6). Across all qualitative data sources, potholes consistently emerged as the most significant and recurrent road damages, reported as the primary cause of poor road user experience and as a challenge that overwhelms existing maintenance capacity. These shortcomings were compounded by resource constraints such as inadequate funding, limited technical skills and shortages of skilled personnel, which hinder systematic and timely interventions. Triangulation of findings ensured validity, and key themes that emerged emphasised the need for modern, technology-driven approaches. Collectively, the study highlighted that while practitioners possess valuable experiential knowledge, there is an urgent need to move towards AI-enabled solutions to improve accuracy, efficiency and evidence-based decision-making in road condition monitoring and maintenance planning.
| TABLE 6: Qualitative data analysis: Categories and sub-categories. |
Discussion of results and findings
The findings demonstrate that integrating machine vision, ESNLP and robotics significantly enhances the accuracy and efficiency of road condition assessment. These results align with previous studies that underscore the importance of AI-based approaches over conventional inspection methods. The outcomes directly address the study objectives by confirming that the combined application of these technologies explains a substantial proportion of the variance in road condition evaluation. Qualitative insights further emphasise the importance of proactive, preventative maintenance strategies and the development of skills within the CoJ Municipality to support technology adoption. This study contributes to the growing body of literature on AI-enabled infrastructure management by demonstrating the potential of intelligent systems to optimise decision-making and resource allocation in municipal road maintenance. Nevertheless, the findings should be interpreted considering certain limitations: the purposive sample size constrains generalisability; contextual factors specific to the CoJ Metropolitan Municipality may not apply elsewhere, and reliance on stakeholder perceptions introduces the possibility of bias despite the use of triangulation.
Recommendations to the City of Johannesburg Metropolitan Municipality
To enhance road maintenance operations, it is recommended that the CoJ Metropolitan Municipality should adopt advanced machine vision technologies to improve the precision, efficiency and cost-effectiveness of road inspections. Qualitative findings from interviews and focus groups strongly highlighted the over-reliance on manual, subjective assessments and highlighted the need for modern, automated systems that can provide real-time and consistent data on road conditions. In addition, the study recommends a broader implementation of preventative maintenance strategies across all regions to reduce the recurrence of severe road damage. These strategies should be supported by region-specific maintenance plans that consider the unique demographic, geographical and usage patterns of each region, as highlighted by participants who reported that certain high traffic areas deteriorate faster than others. Furthermore, the findings indicate the need for capacity building through training and awareness programmes, to equip municipal staff with digital skills and to promote understanding and acceptance of new technologies. Addressing resource constraints and improving coordination between teams were also emphasised in the qualitative results as critical steps to ensure that the adoption of advanced technologies and proactive strategies leads to sustained improvements in road maintenance operations.
Future research
Future research should broaden the geographic scope, employ larger datasets and incorporate longitudinal analyses to examine the long-term impact and cost-effectiveness of AI-driven frameworks in different municipal contexts.
Drawbacks and limitations
Despite its potential, this methodology faces limitations such as the need for high-quality imaging infrastructure, dependence on skilled personnel for model training and interpretation and challenges in integrating data sources. Data privacy, cybersecurity risks and resistance to adopting new technologies in municipalities can also constrain the effectiveness of this framework.
Cost implications
Implementing a machine vision-based Road Condition Assessment framework in South African municipalities involves a substantial initial capital investment estimated between R23 million and R79 million, driven by the acquisition of high-resolution cameras, LiDAR-equipped drones, autonomous robots, computing infrastructure and AI software (Visser et al. 2022). Annual recurring costs are estimated at R8 million to R18 million, covering system maintenance, AI model retraining, software licensing, specialist staff and cybersecurity. Although the upfront costs are significant, the framework enables optimised maintenance planning, reduces emergency repairs and extends road lifespans, leading to long-term cost efficiencies and a typical return on investment within 5 years – 7 years for a large metropolitan municipality.
Conclusion
The study establishes that AI technologies provide a reliable and effective approach to road condition assessment within the CoJ Metropolitan Municipality. By demonstrating the predictive strength of machine vision, ESNLP and robotics, the research highlights their potential to complement existing municipal practices and support data-driven decision making. Importantly, the findings position AI as a strategic tool for advancing infrastructure management while also acknowledging contextual limitations that constrain the generalisability of results. The study therefore contributes to the body of knowledge on AI-enabled public service delivery and offers a foundation for further research exploring its application across diverse municipal environments.
Acknowledgements
Competing interests
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
Authors’ contributions
V.R. conceptualised the study, was contributed to investigation and writing original draft preparation. V.R., T.M. and O.M. developed the methodology and contributed to writing review and editing. All authors read and approved the final article.
Funding information
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Data availability
The data utilised in this investigation were gathered from publicly available sources, including published research articles, as detailed in the manuscript. All relevant data are included within the article and its related materials.
Disclaimer
The views and opinions expressed in this article are those of the authors and are the product of professional research. The article does not necessarily reflect the official policy or position of any affiliated institution, funder, agency of the authors or the publisher. The authors are responsible for this article’s results, findings and content.
References
Abdollahi, A., Pradhan, B., Shukla, N., Chakraborty, S. & Alamri, A., 2020, ‘Deep learning approaches applied to remote sensing datasets for road extraction: A state-of-the-art review’, Remote Sensing 12(9), 1–32. https://doi.org/10.3390/rs12091444
Afridi, M.A., Erlingsson, S. & Sjögren, L., 2023, ‘Municipal street maintenance challenges and management practices in Sweden’, Frontiers in Built Environment 9, 1–18. https://doi.org/10.3389/fbuil.2023.1205235
Ahmed, M.A., Kays, H.I. & Sadri, A.M., 2023, ‘Centrality-based lane interventions in road networks for improved level of service: the case of downtown Boise, Idaho’, Applied Network Science 8(1), 1–18. https://doi.org/10.1007/s41109-023-00532-z
Ahmad, C.F., Al-Sayegh, A.T., Cheema, A., Qayyum, W., Ehtisham, R., Saghir, S. et al., 2024, ‘Classification of different size of potholes based on surface area using convolutional neural network’, Discover Applied Sciences 6(9), 1–24. https://doi.org/10.1007/s42452-024-06207-3
Akinosho, T.D., Oyedele, L.O., Bilal, M., Ajayi, A.O., Delgado, M.D., Akinade, O.O. et al., 2020, ‘Deep learning in the construction industry: A review of present status and future innovations’, Journal of Building Engineering 32, 1–14. https://doi.org/10.1016/j.jobe.2020.101827
Burghardt, T.E., Popp, R., Helmreich, B., Reiter, T., Böhm, G., Pitterle, G. et al., 2021, ‘Visibility of various road markings for machine vision’, Case Studies in Construction Materials 15, 1–15. https://doi.org/10.1016/j.cscm.2021.e00579
Cazzato, D., Cimarelli, C., Sanchez-Lopez, J.L., Voos, H. & Leo, M., 2020, ‘A survey of computer vision methods for 2d object detection from unmanned aerial vehicles’, Journal of Imaging 6(8), 1–38. https://doi.org/10.3390/jimaging6080078
Chen, S.L., Lin, C., Tang, C.W. & Hsieh, H.A., 2022, ‘Evaluation of pavement roughness by the international roughness index for sustainable pavement construction in New Taipei City’, Sustainability 14(12), 1–18. https://doi.org/10.3390/su14126982
Chirisa, I. & Matamanda, A., 2019, ‘Forces shaping urban morphology in Southern Africa Today: unequal interplay among people, practice and policy’, Journal of Urbanism 2019, 1–19. https://doi.org/10.1080/17549175.2019.1626262
City of Johannesburg, 2022, Integrated Annual Report 2021-2022, pp. 82–83, viewed 24 April 2025, from https://joburg.org.za/documents_/Documents/CoJ-IntergratedAnnualReport20212022.pdf.
COJ, 2025, City of Johannesburg Integrated Development Plan, viewed 26 June 2025, from https://www.nra.co.za/uploads/65/SANRAL%20Annual%20Report%202024-25%20%281%29_0.pdf.
CSIR, 2023, Council of scientific and industrial research stormwater management, viewed 26 June 2025, from https://www.csir.co.za/sites/default/files/Documents/Chapter_06_pg%201%20_%206.pdf.
Deni, A., 2020, Application of pavement condition index (PCI) on the assessment of the Kalumata highway section of the City of South Ternate, pp. 1–2, IOP Publishing, Sweden.
Duvenage, W., 2025, ‘Joburg’s traffic lights and roads crisis a symptom of failed leadership and disrespect for residents’, Daily Maverick, pp. 1–2, viewed 10 March 2025, from https://www.dailymaverick.co.za/opinionista/2025-02-21-joburgs-traffic-lights-roads-crisis-a-symptom-of-failed-leadership/.
El Hakea, A. & Fakhr, M., 2023, ‘Recent computer vision applications for pavement distress and condition assessment’, Automation in Construction 146, 104664. https://doi.org/10.1016/j.autcon.2022.104664
Fusch, P. & Ness, L., 2015, ‘Are we there yet? Data saturation in qualitative research’, Qualitative Report 20(9), 1–13. https://doi.org/10.46743/2160-3715/2015.2281
Hancock, T., 2022, ‘Maintenance of ageing road network a top priority – JRA’, Engineering News, pp. 1–3, viewed 25 June 2025, from https://www.engineeringnews.co.za/article/maintenance-of-ageing-road-network-a-top-priority-jra-2022-06-24.
Hennink, M. & Kaiser, B.N., 2022, ‘Sample sizes for saturation in qualitative research: A systematic review of empirical tests’, Social Science & Medicine 292, 1–10. https://doi.org/10.1016/j.socscimed.2021.114523
Hidayat, R. & Wulandari, P., 2022, ‘Structural equation modelling (SEM) in research: Narrative literature review’, Open Access Indonesia Journal of Social Sciences 5(6), 1–6. https://doi.org/10.37275/oaijss.v5i6.141
Huang, J., Han, C., Wei, W. & Zhao, C., 2023, ‘Analysis of long-term maintenance decision for asphalt pavement based on analytic hierarchy process and network level optimization decision’, Electronic Research Archive 31(9). https://doi.org/10.3934/era.2023299
Huang, L.L., Lin, J.D., Huang, W.H., Kuo, C.H., Chiou, Y.S. & Huang, M.Y., 2024, ‘Developing pavement maintenance strategies and implementing management systems’, Infrastructures 9(7), 1–24. https://doi.org/10.3390/infrastructures9070101
Javaid, M., Haleem, A., Singh, R.P., Rab, S. & Suman, R., 2022, ‘Exploring impact and features of machine vision for progressive industry 4.0 culture’, Sensors International 3, 1–11. https://doi.org/10.1016/j.sintl.2021.100132
Jia, X., Woods, M., Martinez, U., Zhu, D. & Huang, B., 2023, ‘Investigation into collection variability of surface crack data for network-level asphalt pavement evaluation’, Transportation Research Record 2677(1), 1–16. https://doi.org/10.1177/03611981221105069
Khan, N.A., Jhanjhi, N.Z., Brohi, S.N., Usmani, R.S.A. & Nayyar, A., 2020, ‘Smart traffic monitoring system using unmanned aerial vehicles (UAVs)’, Computer Communications 157, 1–9. https://doi.org/10.1016/j.comcom.2020.04.049
Kim, S., Anagnostopoulos, G., Barmpounakis, E. & Geroliminis, N., 2023, ‘Visual extensions and anomaly detection in the pNEUMA experiment with a swarm of drones’, Transportation Research Part C: Emerging Technologies 147, 1–25. https://doi.org/10.1016/j.trc.2022.103966
Kim, Y.-M., Kim, Y.-G., Son, S.-Y., Lim, S.-Y., Choi, B.-Y. & Choi, D.-H., 2022, ‘Review of recent automated pothole-detection methods’, Applied Sciences 12(11), 1–15. https://doi.org/10.3390/app12115320
Kitsios, F. & Kamariotou, M., 2021, ‘Artificial intelligence and business strategy towards digital transformation: A research agenda’, Sustainability 13(4), 1–14. https://doi.org/10.3390/su13042025
Lemonakis, P., Kopelias, P. & Karlaftis, P., 2023, ‘Introduction of a new index to assess pavement functional condition’, Environment, Development and Sustainability 26(4), 1–18. https://doi.org/10.1007/s10668-023-04012-3
Lendra, L., Wibowo, M.A. & Hatmoko, J.U.D., 2023, ‘A systematic literature network analysis: Research mapping of international roughness index’, Instrumentation Mesure Métrologie 22(3), 1–15. https://doi.org/10.18280/i2m.220301
Li, W. & Hsu, C.Y., 2022, ‘GeoAI for large-scale image analysis and machine vision: Recent progress of artificial intelligence in geography’, ISPRS International Journal of Geo-Information 11(7), 1–44. https://doi.org/10.3390/ijgi11070385
Lin, C., Loong Chen, S., Tang, C. & Ang Hsieh, H., 2022, ‘A Study on the Roads Flatness Using International Roughness Index in New Taipei City’, Journal of Physics: Conference Series 2160(1), 1–8. https://doi.org/10.1088/1742-6596/2160/1/012035
Marianingsih, S., Widodo, W., Pieter, M.S.S., Manullang, E.V. & Nanlohy, H.Y., 2022, ‘Machine vision for the various road surface type classification based on texture feature’, Journal of Mechanical Engineering Science and Technology (JMEST) 6(1), 1–7. https://doi.org/10.17977/um016v6i12022p040
Moolla, T. & Tetley, S., 2021, The journey towards autonomous full spectrum pavement condition data acquisition and evaluation in south africa: A road less travelled, Southern African Transport Conference 2021, viewed from http://hdl.handle.net/2263/82377.
Moyo, T., Mbatha, S., Aderibigbe, O.-O., Gumbo, T. & Musonda, I., 2022, ‘Assessing spatial variations of traffic congestion using traffic index data in a developing city: Lessons from Johannesburg, South Africa’, Sustainability 14(14), 1–16. https://doi.org/10.3390/su14148809
Patra, S., Middya, A. & Roy, S., 2021, ‘PotSpot: Participatory sensing based monitoring system for pothole detection using deep learning’, Multimedia Tools and Applications 80, 1–24. https://doi.org/10.1007/s11042-021-10874-4
Pavel, M.I., Tan, S.Y. & Abdullah, A., 2022, ‘Vision-based autonomous vehicle systems based on deep learning: A systematic literature review’, Applied Sciences 12(14), 1–51. https://doi.org/10.3390/app12146831
Pereira, P.M. & Vieira, C.S., 2022, ‘A literature review on the use of recycled construction and demolition materials in unbound pavement applications’, Sustainability 14(21), 1–28. https://doi.org/10.3390/su142113918
Peters-Scheepers, 2023, Pothole Patrol, Discovery, p. 105, viewed 10 March 2025, from https://www.jra.org.za/wp-content/uploads/2023/06/JRA_Discovery_Avis_Partnership.pdf.
Pinatt, J.M., Chicati, M.L., Ildefonso, J.S. & Filetti, C.R.G.D., 2020, ‘Evaluation of pavement condition index by different methods: Case study of Maringá, Brazil’, Transportation Research Interdisciplinary Perspectives 4, 1–10. https://doi.org/10.1016/j.trip.2020.100100
Ren, Z., Fang, F., Yan, N. & Wu, Y., 2022, ‘State of the art in defect detection based on machine vision’, International Journal of Precision Engineering and Manufacturing-Green Technology 9(2), 1–72. https://doi.org/10.1007/s40684-021-00343-6
Rose, R., 2024, ‘Johannesburg mayor quits after turbulent 15 months in charge’, Financial Times, pp. 1–2, viewed 10 June 2025, from https://www.ft.com/content/2e3cf9e8-9bf4-4dfe-b675-da53ce0a7837.
Salcedo, E., Jaber, M. & Requena Carrión, J., 2022, ‘A novel road maintenance prioritisation system based on computer vision and crowdsourced reporting’, Journal of Sensor and Actuator Networks 11(1), 1–22. https://doi.org/10.3390/jsan11010015
Sanral, 2024, SANRAL Annual Report 2024/2025, pp. 76–77, viewed 17 June 2025, from https://www.nra.co.za/uploads/65/SANRAL%20Annual%20Report%202024-25%20%281%29_0.pdf.
Schattler, K., Wolters, A. & Zimmerman, K., 2020, Implementing pavement management systems for local agencies: State of the art/state of the practice, pp. 1–99, Illinois Centre for Transportation, viewed from https://www.ideals.illinois.edu/items/45799.
Shah, S., Ghomeshi, H., Vakaj, E., Cooper, E. & Fouad, S., 2023, ‘A review of natural language processing in contact centre automation’, Pattern Analysis and Applications 26(3), 1–24. https://doi.org/10.1007/s10044-023-01182-8
Starke, M. & Geiger, C., 2022, ‘Machine vision based waterlogged area detection for gravel road condition monitoring’, International Journal of Forest Engineering 33(3), 1–6. https://doi.org/10.1080/14942119.2022.2064654
Visser, H., Tetley, S., Lewis, T., Naidoo, K., Moolla, T. & Pillay, J., 2022, ‘A study into the benefit and cost-effectiveness of using state-of-the-art technology for road network level condition assessment’, in A. Akhnoukh, K. Kaloush, M. Elabyad, B. Halleman, N. Erian, S. Enmon, II et al. (eds.), Advances in road infrastructure and mobility, pp. 19–31, Springer International Publishing, Cham.
Wang, Q., Zhou, M., Sabri, M.M.S. & Huang, J., 2022, ‘A comparative study of AI-based International Roughness Index (IRI) prediction models for Jointed Plain Concrete Pavement (JPCP)’, Materials 15(16), 1–16. https://doi.org/10.3390/ma15165605
Wintruff, N.C. & Fernandes, J.L., 2023, ‘A review on life cycle assessment of pavements in Brazil: Evaluating environmental impacts and pavement performance integrating the international roughness index’, Sustainability 15(19), 1–20. https://doi.org/10.3390/su151914373
Yang, X., Zhang, J., Liu, W., Jing, J., Zheng, H. & Xu, W., 2024, ‘Automation in road distress detection, diagnosis and treatment’, Journal of Road Engineering, 4(1), 1–26. https://doi.org/10.1016/j.jreng.2024.01.005
|
Hide
Show all
Email this article (Login required)
