Road maintenance sustainability entails implementing practices that reduce the impact of road accident. while simultaneously ensuring the durability and functionality of the road infrastructure. Pavement distress is a major concern for transportation agencies as it affects the safety and comfort of road users. This paper presents a novel approach for prioritizing pavement distress through the application of image processing techniques and the Analytic Hierarchy Process (AHP). The proposed method involves capturing images of the pavement surface using a high-resolution camera and analyzing them using image processing algorithms. The images are processed to identify different types of pavement distress such as cracks, potholes, and rutting. The severity of each type of distress is then quantified. AHP is utilized to prioritize the identified distress based on its severity and impact on road users. The proposed approach has been tested on real-world pavement images, and the results demonstrate its effectiveness in accurately identifying and prioritizing pavement distress. This method can assist transportation agencies in making informed decisions about maintenance and repair activities, leading to improved road safety and reduced maintenance costs. This automated method achieved an accuracy percentage of about 95%.
Introduction
Roads are of the highest significance in our society since they are the foundation upon which transportation and connection are built. They make it possible for people, products, and services to travel from one place to another, which is essential to the growth of the economy, the integration of society, and general progress. Trade and commerce are both made easier by roads, which also serve to link manufacturing centers with consumer markets and to fuel economic expansion. They make vital services more accessible, which paves the way for people to get access to educational and medical options, as well as career possibilities. In addition, highways foster social connectedness by allowing individuals to pay their loved ones visits, take part in social events, and communicate their cultural experiences with one another.
Roads also help to promote regional integration, which in turn encourages cooperation and commerce across various areas, and they provide a contribution to the handling of emergency situations and natural disasters. In general, roads are very important for creating economic development, strengthening social cohesion, and improving the quality of life for individuals as well as communities. Pavement distress is a significant issue that transportation agencies worldwide face. The pavement condition has a significant impact on the safety and comfort of road users. Cracks, potholes, rutting, and other forms of pavement distress can lead to accidents, increase vehicle wear and tear, and cause discomfort to drivers and passengers. Road traffic crashes are becoming a major concern across the world [1]. Based on the most recent latest World Health Organization data, road accidents kill over 1.25 million population and harm 50 million per year [2]. In addition, pavement distress can also result in higher maintenance costs, reduced road lifespan, and increased traffic congestion, all of which can have significant economic implications. Therefore, identifying and addressing pavement distress is a critical concern for transportation agencies, and effective pavement management strategies can help ensure the safety and comfort of road users while minimizing costs and maximizing pavement lifespan. Any nation's progress in the global economy may be seen in the condition of its highway system. Construction managers are always on the lookout for less expensive and more efficient methods of managing and maintaining pavements to ensure they continue to provide a sufficient level of service despite rising material costs. The assessment of pavement conditions and prioritization of maintenance tasks are integral components in the development of an optimal road maintenance plan.
Anzeige
Transportation agencies utilize a range of techniques and technologies to evaluate pavement conditions, such as visual inspection, pavement surface evaluation, and non-destructive testing methods (Leonardi et al. 2019). Visual inspection is the most commonly used method for identifying pavement distress, as it is relatively inexpensive and provides a quick assessment of the pavement's condition. However, visual inspection is subjective and may not always detect subtle forms of pavement distress. Therefore, pavement surface evaluation and non-destructive testing methods, such as ground-penetrating radar and infrared thermography, can provide more detailed and objective information on pavement conditions. Pavement cracks as quickly as possible. Several research studies have been conducted to develop a method for automated identification of pavement cracks that overcomes the limitations of the manual technique(Burke and Segrin 2014). Therefore, researchers have been working on developing automated methods for identifying pavement cracks, which can be more efficient and accurate. Several research studies have been conducted on the automated identification of pavement cracks, using various techniques such as image processing, machine learning, and artificial intelligence. These studies have demonstrated promising results in terms of accuracy and efficiency. One study by Hou et al. (2020) used deep learning techniques to develop an automated crack detection system. The system was trained on a large dataset of pavement images and was able to accurately identify different types of cracks, including longitudinal, transverse, and alligator cracks, with an accuracy of over 90%.
Another study by Wang et al. (2019) developed a crack detection method based on thermal imaging. The method used a convolutional neural network (CNN) to extract features from thermal images and was able to accurately detect cracks with a detection rate of over 95%. In addition, several studies have focused on developing automated methods for identifying and classifying pavement distresses, where it doesn't only specify cracks but also other types of damage such as potholes and rutting (Tian et al. 2021). For example, a study by Li et al. (2020) used a machine learning approach to classify different types of pavement distresses, achieving an accuracy of over 90%. Image processing has been widely used in the detection and analysis of pavement cracks. The technique involves the acquisition of digital images of pavement surfaces, followed by the application of various algorithms to extract and analyze crack features (Hsieh and Tsai 2020).Several studies have investigated the use of image-processing techniques for pavement crack detection, classification, and quantification. For example, a study by Chen et al. (2018) proposed a method for detecting and classifying pavement cracks using a convolutional neural network (CNN) and achieved an accuracy of over 90% in crack detection [3].
Another study by Oumaamar et al. (2019) developed an automated system for pavement crack detection and quantification using image processing. The system was able to accurately detect and measure the length, width, and spatial distribution of cracks, providing valuable information for pavement maintenance and repair.
In addition, some studies have explored the use of advanced image processing techniques, such as 3D imaging and thermal imaging, for pavement crack analysis. For instance, a study by Wu et al. (2019) proposed a method for 3D crack detection using a stereo vision system, which was able to accurately capture crack depth and shape information.
Anzeige
Overall, these studies demonstrate the potential of image processing techniques for pavement crack analysis, offering advantages such as high accuracy, efficiency, and non-destructive evaluation. As technology continues to advance, these methods will likely become more sophisticated and widely applied in the field of pavement engineering. Developed an image processing technique designed specifically for the fast assessment of cracking in the road surface [4].
Numerous studies have been conducted to prioritize pavement cracks through the utilization of the Analytic Hierarchy Process (AHP). This method involves a systematic approach to decision-making that utilizes a hierarchical structure to develop a prioritization scheme based on multiple criteria. For example, a study by Chen et al. (2017) used AHP to prioritize pavement cracks based on six criteria: crack width, crack length, crack area, traffic volume, road network importance, and economic cost. The study found that crack width was the most important criterion, followed by crack length and crack area. Another study by Wang et al. (2018) used AHP to prioritize pavement cracks based on four criteria: traffic volume, crack severity, road network importance, and economic cost. The study found that crack severity was the most important criterion, followed by traffic volume and road network importance.
In addition, some studies have combined AHP with other methods, such as fuzzy logic and GIS, to improve the accuracy and efficiency of pavement crack prioritization. For instance, a study by Aghayari et al. (2017) used a fuzzy AHP-GIS approach to prioritize pavement cracks based on six criteria: crack width, crack length, crack area, traffic volume, road network importance, and road surface condition. The study found that crack width was the most important criterion, followed by crack length and traffic volume.
Despite the significant contributions of formation studies to the field of pavement maintenance, their limitations prevent them from effectively evaluating pavement conditions and prioritizing maintenance tasks for an optimal maintenance plan. To address this issue, this paper integrates image processing for the identification of pavement cracks with Multi-Criteria Decision Making (MCDM) to prioritize pavement crack maintenance. This approach enables a more comprehensive assessment of pavement conditions and allows for the identification of critical areas that require immediate attention.
Road distress classification
Road distress could be categorized according to its form, breadth, and depth. Distress should be identified as soon as possible before it presents itself. Become a more severe issue the road hazards were divided into five categories groups: Cracking, Patching, potholes, Surface deformation, Surface defects and Miscellaneous distresses as shown in Figs. 1, 2, 3 and 4. The four main types of fractures detected in early pavement degradation are transverse cracks, longitudinal cracks, block cracks, and alligator cracks.
Fig. 1
Classification of road cracks
Fig. 2
Road patching and potholes
Fig. 3
Surface deformation
Fig. 4
Classification of road defects
×
×
×
×
Cracking has a direct impact on pavement performance and the most prevalent cause of pavement deterioration is cracking [5]. There are many reasons for road distress cracks such as; the sudden rise in traffic loads is a key cause of cracking, particularly on new roads when the design is predicated on low traffic. After a high-quality road has been built, drivers will naturally gravitate towards it. The fatigue failure is hastened because the temperature swings between 50 °C and sub-freezing levels cause bleeding and cracking. Edge failures occur when inadequate shoulders are provided, Corrugation at the surface and an increase in unevenness, and the provision of poor clayey subgrade, water seeps through the pavement from the sides and the top during rainy seasons due to inadequate drainage.
When the bituminous layer is openly graded, the top layer might separate from the rest, creating a hazardous situation, and pavement collapse may also occur if bitumen and bituminous mixtures are not kept at the correct temperature. Bitumen loses its adhesive qualities if heated too much. When the bituminous mix’s temperature is too low, improper compaction occurs, and the result is longitudinal corrugations.
Automated versus manual pavement cracks detection
There are two methods for collecting pavement distress data manual and automated data collection. First, manual data gathering is an empirically manual-based inspection technique to detect pavement defects or distresses to conduct cost-effective and streamlined pavement condition inspections [6]. This method utilized the human eye to identify and quantify various forms of cracking and spalling in any quantity, at any collecting pace, and in every weather condition [7]. The changeling of the observed pavement failures and the accompanying pavement distress were obtained using GPS.
Second, automated data collection in which the development and deployment of computer vision algorithms for pavement engineering systems have expanded at an exponential rate [8]. Significant scientific effort has gone into creating automated and semi-automatic processes for pavement testing and evaluation. Strategies for non-destructive examination, like image processing (IP) [9]. While several attempts have been made to acquire pavement crack data mechanically, new methodologies are required to assess these automated crack monitoring systems under diverse conditions [10]. Expenses, speed of processing, repeatability, precision, objective and precise identification or analysis of these cracks, and cost reduction are all critical duties in this type of system. The kind, intensity, and amount of road surface cracking are important factors in determining the quality of road pavement [11]. In the process of locating and evaluating cracks in a paved surface, two of the most common methods of detection are automated and human inspection.
Automated crack detection is the process of automatically locating and analyzing fractures in road surfaces by the use of sophisticated technology and algorithmic processes. This strategy processes vast amounts of data by using methods such as image processing, computer vision, and machine learning in order to find fractures depending on certain criteria that have been established. Automated crack detection provides a number of benefits, including enhanced efficiency, shortened processing times, and a decreased likelihood of errors caused by human intervention. It makes it possible to conduct an examination of huge road networks in a short amount of time and delivers findings that are consistent and objective. However, automated crack detection may have difficulties in coping with complicated fracture patterns, fluctuations in lighting conditions, and the need for frequent calibration and maintenance of the detection systems. These difficulties may arise because of the necessity for regular calibration and maintenance of the detection systems.
On the other hand, manual crack detection is a method that identifies cracks in pavements by visual examination and the use of human skill. Cracks in the road surface are physically inspected by trained inspectors or engineers, who then visually detect the cracks and evaluate both their severity and their extent. Manual detection makes it possible to conduct a thorough investigation of cracks, including an analysis of their dimensions (such as width, length, and kind), as well as their features. In addition to this, it allows inspectors to take into account the surrounding circumstances and to form subjective opinions based on their own experiences. Manual crack detection, on the other hand, requires a lot of work, takes a long time, and might be vulnerable to changes in competence and subjectivity on the part of the inspector. It is possible that it will be difficult to discover tiny fractures or cracks in their early stages, particularly if they are not clearly apparent to the human eye.
In the real world, it is common practice to use a mix of automatic and human detection methods in order to perform the most accurate crack assessment possible. While automated methods may be used for large-scale screening and the preliminary detection of cracks, human inspection is the method of choice for comprehensive analysis, verification, and validation. This hybrid strategy takes use of the efficacy and impartiality of automated detection, while at the same time, employing human knowledge for correct interpretation and decision-making. The management of cracks in road infrastructure may be made more efficient and effective by the combination of automated and manual techniques of crack identification, which can be integrated into the process.
Inventory of Pavement Distress and Failure Index due to gathering comprehensive pavement condition data is time-consuming, Innovative ways for quick data collecting are becoming increasingly popular among highway authorities with limited PMS budgets that creates erratic and variable results. Consequently, it subjects the inspectors to hazardous roadway working circumstances. Non-Destructive Test (NDT) and Destructive Test (DT) are both expensive and time-demanding [12]. To avoid the limits of the terms of the visual evaluation process, several efforts have been taken to build semi-automated and automatic processes [13, 14] therefore automation of pavement cracks is the optimal method than manual method as shown in Table 1.
Table 1
Comparison between manual and automation (image processing) methods
Manual detection method
Automatic detection method
Expensive and time-consuming
Less expensive and fast
Labor intensive
Very minimal labor is needed
Hazardous
Safe
Data sampling
100% survey
Subjective
Objective
Difficult to manage
Integrate with a management system
Repeatability is low
Proven to be much better
Utilizing image processing
Edge detection is an algorithm used in image processing [15]. For this reason, digital image processing often begins with edge detection as one of the initial stages in segmentation, with the ultimate goal of displaying the objects within the image [15‐18]. To determine where an item begins and ends in a merged picture, edge detection is used. All items in the image could be located, and their most basic characteristics like area, shape, and size could be quantified if only the image's edge lines could be correctly identified [19, 20].
The pixel intensity of an image transitions from low to high, or vice versa, at the image's edge, a phenomenon known as the image edge [19]. There are several algorithms available for use in edge detection, including the Sobel, Canny, Prewittfrei-Chenen, and Laplacian [21, 22] algorithms. In this manuscript using the canny ala algorithm, Edge as defined by this study's astute algorithm, is a sharp, localized variation in the intensity of a gray scale [21]. Image objects may be categorized more easily and more in-depth analysis can be performed if the edges between them are identified [23‐25]. The crack inspection on the road is very crucial in determining the condition of the route that is utilized due to the impact on the safe operation of the street drivers, cracks on roads could be considered seriously by fixing the road, and this research makes utilize the threshold function in the proposal, which is a parameter that establishes the degree of image clarity of edge detection outcomes done.
Image processing methods based on a camera
This section provides a short overview of processing methods based on camera images for the identification of fractures in engineering structures. Numerous publications are evaluated here using camera input.
Using digital image processing technologies, Yiyang et al. devised a crack-detecting technique. It was able to get the details about the crack picture by preprocessing, image segmentation, and feature extraction [26]. After the input picture was smoothed and approved, the Threshold technique of segmentation was utilized [26]. The size and perimeter of the roundness index have been determined to evaluate their presentation. The existence of the picture fracture was then assessed by comparison that has assessed whether or not the break is visible in the picture. Although many of the commercially available cameras rely on Whi processing methods rely only on the pre-processing stage, someplace more emphasis is on the integration algorithm, which is where the feature extraction occurs. Adhikari et al. [27] created a model to express the flaws mathematically.
The measurement and identification of cracks, a neural network, and a model for visualizing the data in three dimensions all make up their integration model. Brown and Lowe's [28] feature-based registration picture stitching approach has been used. To get the broken pieces back, they employed a skeletonization technique. Crack identification was dependent on the assessment of the crack quantification model in terms of both breadth and length. The effectiveness of an image processing method might be affected by the filters used in it. A method based on Gabor filtering was presented by Salman et al. [29] to automatically detect fractures in digital photographs. The high potential Gabor filter allows for the identification of cracks in several directions. The Gabor filter shows great promise as a method for detecting cracks in several directions. Gabor filter image analysis has a one-to-one correlation with human visual perception. Filtering allows for the detection of fractures aligned in various directions. Their suggested approach boasts a 95% accuracy rate for detection. In this manuscript was used the y filter that gave accuracy more than Gabor filter the and others [30, 31].
Canny edge detection
The Canny edge detector was developed by John F. Canny in 1986 while he was a graduate student at MIT. The goal of the algorithm was to find the optimal way to detect edges in an image while minimizing noise and false detections. The well-known image processing method known as canny edge detection is used in order to locate and extract edges from digital photographs, it has remained one of the most extensively used edge detection techniques ever since.
The Canny edge detection method is comprised of a few different stages that work together to properly identify edges while simultaneously reducing the amount of noise and erroneous responses. It starts out by applying Gaussian smoothing to the picture in order to decrease the amount of noise, and then it moves on to compute the gradient magnitude and direction. After that, a technique known as non-maximum suppression is used to reduce the thickness of the edges and guarantee that just the sharpest, most distinct edges are kept. The last step is to utilize a hysteresis thresholding strategy, which determines the edges of the final set based on high and low threshold values. Canny edge detection is appreciated for its ability to create accurate and well-defined edges, which makes it a useful tool in different applications including object identification, picture segmentation, and feature extraction in computer vision and image processing jobs. Canny edge detection is valued for its ability to produce precise and well-defined edges. Canny edge detection is valued for its ability to produce exact and well-defined edges. Canny's approach involved several steps, including smoothing the image with a Gaussian filter, calculating the gradient magnitude and direction, non-maximum suppression to thin out edges, and hysteresis thresholding to determine which edges were strong enough to be considered true edges.
The Canny edge detector quickly became one of the most widely used edge detection algorithms due to its effectiveness and robustness. It has been applied in a wide range of fields, including computer vision, robotics, medical imaging, and more.
Over the years, there have been various modifications and improvements made to the original Canny algorithm. These include adaptive threshold techniques, multi-scale approaches for detecting edges at different levels of detail, and deep learning-based methods for edge detection. However, the basic principles behind the Canny edge detector remain relevant today as a fundamental tool for image processing. To identify edges concerning efficiency, the Canny Edge Detector is among the most widely used image processing tools. When it comes to edge detection, the Canny edge detector is almost universally accepted as the gold standard. Based on Canny analysis, the edge detection issue is best seen as an optimization problem in signal processing, and hence an objective function is constructed to be optimized. Canny discovered several approximations and optimizations for the edge-searching issue, even though the answer was a complicated exponential function. Canny edge detection. A canny algorithm examines a small enough margin of error to provide an optimum edge image. There are five primary stages to the Canny algorithm. Edge noise can be filtered out with the help of a high threshold (TH) and a low threshold (TL).
Pixels with a low gradient value to make up for those erroneous responses, while keeping the pixels at the edges where the gradient value is high. The input image's content is used to empirically determine the values for the two thresholds.
Five steps were followed in canny edge detection:
Input: the image of a pavement fracture:
1.
Gaussian filter applied to a smooth image;
2.
Choose the image's magnitude gradients;
3.
Eliminate spurious edge detection responses using non-maximum suppression;
4.
Using the Canny approach (TH and TL values), determine the threshold;
5.
Using the suggested work, determine new TH and TL values;
6.
Hysteresis on the track edge (removing any frail edges);
Output: Creation of an edge image: as shown in Figures 6, 7, 8, 9, 10, 11, 12, 13 and 14.
Methodology
The objective of image processing is to identify and extract the cracks of the pavement from images that are deteriorated. Extraneous features with pixel intensities greater than the average pixel intensity in the picture are removed during preprocessing. The procedure involves pushing to the background any pixels depicting paint striping or surface patterns that are brighter than the typical background gray level. This example shows how pavement image processing extract and identify pavement cracks.
MATLAB, a comprehensive programming software then uses MCDM methods in priority was used in this research as shown in Figure 5.
Fig. 5
Schematic diagram of automated road cracks and priority in maintenance
×
Image acquisition
The process of acquiring digital pictures from different devices, such as digital cameras, scanners, or other image sensors, is referred to as image acquisition. The optical picture is changed into a digital signal that may be stored and processed by a computer or other electronic device as part of the process. Image acquisition in digital photography entails taking a picture using a digital camera in taking images with the specification:
108 MP primary sensor with f/1.8 aperture and OIS (optical image stabilization)
12 MP ultra-wide sensor with f/2.2 aperture and 120° field of view
12 MP periscope telephoto sensor with f/3.0 aperture, 5 × optical zoom, and OIS.
Noise removal
Noise reduction removes random changes and distortions from a digital signal or picture. Noise might result from imaging sensor limits, compression, or environmental issues like bad lighting. Noise might appear as unpredictable brightness, color, graininess, or blurring. Noise reduction methods remove these artifacts while keeping the image’s important elements. Simple filtering removes low-frequency noise, but more complex algorithms utilize statistical analysis and machine learning to detect and eliminate particular noise types. Median, Gaussian, wavelet, and picture in painting remove noise. The noise kind, severity, and picture application determine the noise reduction approach.
Edge detection
Edge detection in digital image processing locates the borders of sections of an image with various properties, such as color, intensity, or texture. Edge detection is a basic image processing approach that tries to discover and extract the boundaries between various objects or areas included within an image. It does this by analyzing the picture in question to locate and pinpoint these boundaries. The tasks of computer vision, image analysis, and pattern recognition all benefit greatly from its incorporation. Finding substantial variations in intensity or color within an image is the objective of edge detection. This is because significant changes in intensity or color often correlate to the borders of objects or the transitions between distinct textures or areas. Edge detection methods give crucial information for further image analysis and processing processes by locating and emphasizing edges that have been located in an image. There is a wide range of complexity and methodology that can be found in edge detection algorithms. These algorithms may range from simple gradient-based methods such as the Sobel or Prewitt operators to more complicated approaches such as the Canny edge detection algorithm.
Object identification, picture segmentation, feature extraction, and image enhancement are examples of common uses. The edges of a picture are where the intensity of the pixel values suddenly changes, suggesting a change in the underlying object or surface. Edge detection uses mathematical algorithms or filters to recognize sudden intensity changes and indicate object or area boundaries. Sobel, Canny, and LoG operators are typical edge detection methods. These techniques emphasize edges while reducing noise and low-level features by convolving the picture using a kernel or filter.
Computer vision, image analysis, and machine learning use edge detection for object recognition, picture segmentation, and feature extraction. These applications employ it as a pre-processing step to minimize data and extract useful characteristics for analysis or categorization.
Morphological operation
Morphological processes alter visual objects' shapes and structures. These techniques apply mathematical morphology theory to binary or grayscale pictures. Morphological procedures extract features, improve pictures, and eliminate noise and other artifacts. Erosion and dilation make up morphology. Erosion shrinks picture objects by removing pixels from their edges. This process is used to eliminate tiny items and smooth bigger ones. Dilation expands things by adding pixels to their borders. This procedure fills gaps and enlarges things. Opening and shutting are frequent morphological processes. Opening removes tiny items and noise from images by combining erosion and dilatation. Closing uses dilatation and erosion to patch gaps or reattach fractured pieces.
Morphological procedures may segment binary or grayscale pictures, extract features, and recognize patterns. Application and image properties determine operation and settings. Computer vision, image analysis, and machine learning use edge detection for object recognition, picture segmentation, and feature extraction. These applications employ it as a pre-processing step to minimize data and extract useful characteristics for analysis or categorization.
Classification
Road pavement cracks are classified by their features. Age, weathering, traffic, and environmental conditions may cause pavement fissures. Pavement cracks include longitudinal, transverse, block, alligator, and others. Pavement management benefits from crack categorization. By classifying cracks by severity and road safety effect, maintenance, and repair may be prioritized. Large transverse cracks may need rapid care, but minor longitudinal fractures may not. Pavement crack categorization helps enhance pavement care programs, making roads safer and more dependable. The term "pavement classification" refers to the process of categorizing and classifying various kinds of road surfaces based on their features and attributes. Pavement classification may also be used as a noun. It is a crucial component of both the engineering of transportation systems and the administration of physical infrastructure.
The categorization of pavement takes into consideration a variety of characteristics, including the material, composition, and design of the pavement, as well as its thickness, condition, and performance. Pavements may be categorized so that transportation authorities and engineers have a better understanding of their behavior, can more accurately forecast how well they will function, and can make more educated judgments about the activities of maintenance, rehabilitation, and building. Road surfaces may be evaluated for their structural capability, durability, and potential service life with the use of pavement categorization systems. In addition to this, they are helpful in determining which materials and design criteria are best suited for new pavement construction projects. Pavement categorization also enables efficient asset management, the priority of maintenance operations, and the optimization of resources for the administration of road networks. In general, the categorization of the pavement is an important factor in ensuring that the road infrastructure is safe, efficient, and sustainable.
Quantification
Quantifying and analyzing road pavement cracks is called quantification. Age, weathering, traffic loads, and environmental conditions may cause pavement cracks of varied sizes, shapes, and severity. Pavement crack quantification entails evaluating fracture damage and estimating its influence on structural integrity.
Edge detection, threshold, and segmentation can measure pavement fractures. These methods can extract pavement crack length, breadth, area, and orientation from photographs. Based on these variables, regression analysis can predict fracture severity.
Pavement maintenance may benefit from crack quantification. Quantifying fracture damage helps prioritize road safety-related maintenance and repair. Larger fractures may need rapid care. Quantifying pavement cracks may also track pavement quality over time and evaluate repair efforts to reduce crack formation. Quantifying pavement cracks improves pavement repair programs, making roads safer and more dependable.
Analytic hierarchy process
It is a decision-making method that helps individuals or groups to prioritize and make complex decisions by breaking them down into smaller, more manageable parts. AHP involves creating a hierarchy of criteria and alternatives, assigning weights to each criterion based on its importance, and then evaluating each alternative against the criteria to determine the best option. It was developed by Thomas Saaty in the 1970s and has since been widely used in fields such as business, engineering, and healthcare.
Analytic hierarchy process (AHP) steps
a.
Define the problem: the first step in the AHP process is to clearly define the problem and identify the decision that needs to be made.
b.
Identify criteria: next, you need to identify the criteria that will be used to evaluate the alternatives. These criteria should be relevant, measurable, and mutually exclusive.
c.
Establish a hierarchy: once you have identified the criteria, you need to establish a hierarchy that shows how they are related to each other according to relative importance as shown in Table 2. This hierarchy will help you determine which criteria are more important than others.
d.
Assign weights: after establishing the hierarchy, you need to assign weights to each criterion based on its relative importance. These weights should add up to 100%
e.
Evaluate alternatives: with the criteria and their weights established, you can now evaluate each alternative against each criterion.
f.
Calculate scores: after evaluating all of the alternatives against all of the criteria, you can calculate scores for each alternative based on how well it performs against each criterion.
g.
Analyze results: finally, you need to analyze the results and make a decision based on your findings. This may involve comparing scores for different alternatives or considering other factors such as cost or feasibility.
Table 2
Relative Importance Scale
Definition
Relative importance intensity
Equal importance
1
Moderate importance
3
Strong importance
5
Very importance
7
Extra strong
9
There are intermediate values
2,4,6,8
The reciprocal value of rij is 1/rij, and the judgment value of the significance of the elements I and j is rij
The reciprocal value
Ranking and consistency check
The relative relevance of each element of one layer to the element of the above layer may be retrieved after generating the comparison matrix. The eigenvector corresponding to the primary eigenvalue of the judgment matrix can be normalized to calculate importance.
The developed judgment matrices are used to quantify the decision-making process. When a large number of people are involved. When pairwise comparisons are made, discrepancies will occur.
If the consistency ratio is smaller than 0.10, then the pairwise comparisons have a fair level of consistency. The values of the ratio are suggestive of inconsistent judgments if it is more than 0.10 [32].
The goal of the matrix consistency check is to guarantee that the assessment is consistent and that each judgment is sensible to avoid any contradicting findings. In practice, perfect constancy is unusual. The judgment matrix is regarded as sufficiently consistent if the associated consistency ratio (CR) is less than 10%. To begin, Eq. (1) is used to compute a Consistency Index (CI) based on the maximal eigenvalue max and get RI according to several criteria (n) as shown in Table 3:
Table 3
The random consistency index (RI) values
N
1
2
3
4
5
6
7
8
9
RI
0
0
0.58
0.9
1.12
1.26
1.36
1.41
1.45
$$CI=\frac{{\lambda }_{{\text{max}}}-n}{n-1}$$
(1)
Then, as demonstrated in, divide the CI by the Random Consistency Index (RI) to get the CR as shown in Eq. (2).
$$CR = \frac{CI}{{RI}}$$
(2)
Checking for synthesis and consistency
Calculating the ranking weights of the relative relevance of all elements of a specific layer to the top layer is the final hierarchy priority ranking. The decision maker must generate N judgment matrices (one for each criterion) of order M * M (form factors) and one judgment matrix of order N * N for a situation with M choices and N criteria as described in this work (for n criteria). The ultimate priority of the options, designated by WP1, WP2, and WPi, in terms of all criteria combined, is calculated using Eq. (3)
where CI(j) is the criterion j's consistency index CI, and RI(j) is the criterion j’s average random consistency index RI.
Case study
A real-life case study of a highway project is analyzed to illustrate the use of the proposed method and demonstrate its capabilities. The case study focuses on the Western Upper Desert Road development project spanning from Cairo to Minya. The significant asphalt defects in this project have negatively impacted the safety and quality conditions of the road. it should be noted that this highway project is part of the National Roads Project, encompassing a total length of 7,000 km and estimated costs of 6 billion dollars. The proposed method was utilized to prioritize pavement distress of the project through the application of image processing techniques and the Analytic Hierarchy Process (AHP). First image processing techniques were utilized to accurately identify and quantify the various types of defects. Additionally, AHP was employed to prioritize and classify the repairs needed for the identified defects. In this study, a section of the road measuring 1 km was investigated to detect faults and evaluate the efficacy of the proposed method. The analysis was conducted under three assumptions: (1) the imposition of a fixed time constraint, (2) the imposition of a fixed budget constraint, and (3) the absence of any fixed time or budget constraints.
Result discussion
The suggested method has been implemented in MATLAB, which was applied to a section with a total length of 1 km the performance, simulation, and crack quantities results of the developed algorithm are reported in this part as shown in Table 4.We have taken into account several types of pavement pictures in our implementation, such as transverse, longitudinal, or numerous cracks in a pavement image as shown in Figs. 6, 7, 8, 9, 10, 11, 12, 13 and 14.
Table 4
Pavement cracks quantities
Roads / Distress (Km)
Potholes (cm2)
Rutting (cm2)
Delamination (cm2)
Longitudinal cracks (cm)
Transverse cracks (cm)
Alligator cracks (cm2)
Polished aggregate (cm2)
Bleeding (cm2)
0–0.15
454.5
0
3822.6
428
0
0
1415.5
0
0.15–0.3
0
6300
0
0
1064
7792
14,000
0
0.3–0.45
0
4326
0
1300
0
4898
0
4680
0.45–0.6
0
0
2091
0
420
0
3462
0
0.6–0.75
10,000
6199
0
989.2
0
15,278
0
0
0.75–1
0
7850
7718.2
487.8
0
0
5793.4
0
Fig. 6
Pavement cracks detection by image processing
Fig. 7
Pavement cracks detection by image processing
Fig. 8
Pavement cracks detection by image processing
Fig. 9
Pavement cracks detection by image processing
Fig. 10
Pavement cracks detection by image processing
Fig. 11
Pavement cracks detection by image processing
Fig. 12
Pavement cracks detection by image processing
Fig. 13
Pavement cracks detection by image processing
Fig. 14
Pavement cracks detection by image processing
×
×
×
×
×
×
×
×
×
By evaluating the outcome of each assumption, the effectiveness and versatility of the proposed method were assessed of the varying project constraints. The results of the simulation and quantities of cracks obtained from the developed algorithm are reported in this section, as presented in Tables 4 and 5. The Results are divided into two sections, the first is the quantification of defects in the road section as shown in Table 4, the proposed method implementation utilized 38 images. These images include the diverse crack patterns observed in the pavement such as transverse, longitudinal, or numerous cracks in the pavement as shown in Figs. 6, 7, 8, 9, 10, 11, 12, 13 and 14. The first is the quantification of defects in the road section These images show the initial pavement image with cracks, the upgraded image, the morphologically filtered image, the Canny edge detection, and the final linked crack feature. After enhancement, the image appears to have reduced noise and significantly increased contrast.
Table 5
Criteria weight of pavement cracks
Factor weight
W1
W2
W3
W4
W5
W6
W7
W8
λ max
CI
RI
CR
WC
0.3872
0.2113
0.1489
0.1026
0.0585
0.0414
0.0295
0.0205
8.885
0.12647
1.41
0.089692
Quantification of pavement cracks
The measurement of the quantity, length, and severity of cracks on a particular pavement surface is the quantification of pavement cracks. This data is crucial for assessing the state of the pavement and creating a strategy for maintenance or reconstruction. Automated crack detection systems employ cutting-edge tools like artificial intelligence and digital image processing. The quantification of pavement crack was calculated by automated method from images as shown in Table 4. Image Processing was used to improve the images and extract the defects and their severity by using canny edge detection and morphological operation in the following images:
Priority of pavement maintenance using Analytic hierarchy process (AHP)
AHP is a decision-making method commonly used in infrastructure management, including prioritizing pavement cracks for maintenance and repair. AHP involves breaking down a complex decision problem into a hierarchy of criteria and alternatives, and then using pairwise comparisons to establish the relative importance of each criterion and alternative.
Several studies have used AHP to prioritize pavement cracks for maintenance and repair. It is worth noting that the relative importance of each criterion which are potholes, rutting, delamination, alligator, polished aggregates, longitudinal, transverse, and bleeding may vary depending on the specific context and objectives of the decision-making process. Therefore, it is important to carefully consider the criteria and weights used in the decision-making process to ensure that the prioritization of pavement cracks is appropriately tailored to the specific needs and goals of the organization responsible for pavement maintenance and repair as shown in Table 5.
Road maintenance plans according to time and budget
In network pavement maintenance planning, the complete ranking indicates the urgency and a road stretch. The cost of pavement maintenance for patching per m2 is $100 and $10 for cracks [33]. The priority ranking findings showed the relevance and urgency of each road stretch in pavement maintenance that has a high ranking which requires maintenance need than other (AHP) method. Three cases were considered in the research; case (1) where a fixed time is available for the repair work, case (2) where a fixed budget is available, and finally case (3) if there is no limit for time or budget.
Case (1) fixed time
Fixed time scheduling can refer to a meeting that is scheduled to begin and conclude at a given time, regardless of other circumstances like the completion of a task or the presence of participants. In planning, a deadline that has been predetermined and cannot be adjusted is referred to as a fixed time. A deadline or other time restriction is placed on the decision-making process. Typically refers to a predetermined amount of time that is set in advance and does not change. This can refer to a variety of contexts, such as scheduling, planning, or decision-making. 14 days had been allocated to repair per kilometer of the road as shown in Table 7.
Table 7
Maintenance of pavement cracks with fixed time
No
Crack type
Quantity
Severity
Time of maintenance per unit area (unit/day)
Cost per unit area (unit/$)
Priority
1
Potholes (m2)
1.0455
high
3–4 days
$100
1
2
Rutting (m2)
2.4675
high
3–4 days
$100
2
3
Delamination (m2)
1.3632
Low
3–4 days
$100
3
4
Longitudinal cracks (m)
32.05
Low
Not compete
Not compete
4
5
Transverse cracks (m)
14.84
Low
Not compete
Not compete
5
6
Alligator cracks (m2)
2.7968
high
Not complete
Not complete
6
7
Polished aggregate (m2)
2.4671
medium
Not complete
Not complete
7
8
Bleeding (m2)
0.4680
Low
Not complete
Not complete
8
Total
12 days
$ 487.3
Case (2) Fixed budget
Fixed budget is a set amount of money that is provided for a certain goal, undertaking, or time period and that cannot be altered or exceeded without prior consent. The fixed budget may be based on past performance, forecasts, or other elements pertinent to the budget's goal. Additionally, rigid, set budgets may not account for unforeseen expenses or changes in circumstances. The fixed budget may be compared to actual expenses or revenues in specific circumstances, and modifications may be made as needed. All things considered, fixed budgets may be a helpful tool for managing money, but they must be meticulously developed and managed to ensure that they fulfill the goals of the company. $1,000 had been allocated to repair per kilometer of the road as shown in Table 8.
Table 8
Maintenance of pavement cracks with fixed budget
No
Crack type
Quantity
Severity
Time of maintenance per unit area (unit/day)
Cost per unit area (unit/$)
Priority
1
Potholes (m2)
1.0455
high
3–4 days
$100
1
2
Rutting (m2)
2.4675
high
3–4 days
$100
2
3
Delamination (m2)
1.3632
Low
3–4 days
$100
3
4
Longitudinal cracks (m)
32.05
Low
2–3 days
$10
4
5
Transverse cracks (m)
14.84
Low
2–3 days
$10
5
6
Alligator cracks (m2)
2.7968
high
Not complete
Not complete
6
7
Polished aggregate (m2)
2.4671
medium
Not complete
Not complete
7
8
Bleeding (m2)
0.4680
Low
Not complete
Not complete
8
Total
18 days
$947.6
Case (3) no fixed time and budget
Projects or tasks may sometimes lack a set deadline or budget, which means that the time and resources needed to execute them are not defined or constrained.
This may be true for initiatives that are exploratory in character, whose final objective is still being completely defined, or whose scope may vary in response to new findings or information. For instance, scientific research projects could not have a set duration or cost since the project's parameters might alter in response to fresh information. To make sure that progress is being made and the project is going ahead when there is no set schedule or budget, it is crucial to set clear objectives and milestones. Additionally, it is crucial to periodically review the project and revise the timeframe and objectives as necessary in light of fresh information or evolving conditions. In these kinds of circumstances, efficient project management strategies, like agile project management, may be very helpful as shown in Table 9.
Table 9
Pavement cracks quantity and cost with no fixed time and budget
No
Crack type
Quantity
Severity
Time of maintenance per unit area (unit/day)
Cost per unit area (unit/$)
Priority
1
Potholes (m2)
1.0455
high
3–4 days
$100
1
2
Rutting (m2)
2.4675
high
3–4 days
$100
2
3
Delamination (m2)
1.3632
Low
3–4 days
$100
3
4
Longitudinal cracks (m)
32.05
Low
2–3 days
$10
4
5
Transverse cracks (m)
14.84
Low
2–3 days
$10
5
6
Alligator cracks (m2)
2.7968
high
1–2 days
$100
6
7
Polished aggregate (m2)
2.4671
medium
1–2 days
$100
7
8
Bleeding (m2)
0.4680
Low
1–2 days
$100
8
Total
24 days
$1532
Applicability of model
The advantages of incorporating image processing and Multi-Criteria Decision Making (MCDM) techniques in the sustainability of road maintenance are as follows:
a.
Accurate assessment: image processing allows for the accurate assessment of road conditions by analyzing images captured by cameras or drones. This technology can identify various road defects such as cracks, potholes, and pavement distress, enabling timely maintenance interventions.
b.
Cost and time efficiency: image processing combined with MCDM techniques can streamline the road maintenance process, leading to cost and time savings. By automating the assessment and decision-making process, road authorities can prioritize maintenance activities based on objective criteria, optimizing resource allocation and reducing delays.
c.
Increased safety: timely detection and repair of road defects through image processing enhance road safety. By addressing issues such as potholes or pavement cracks promptly, the risk of accidents and damage to vehicles is minimized, ensuring safer travel conditions for motorists.
d.
Enhanced sustainability: integrating image processing and MCDM in road maintenance supports sustainability goals. By accurately identifying maintenance needs, resources can be utilized efficiently, reducing unnecessary repairs and material waste. Optimized maintenance planning also extends the lifespan of road infrastructure, reducing the need for frequent reconstruction and minimizing the environmental impact.
e.
Data-driven decision making: image processing provides valuable data on road conditions, which can be combined with MCDM techniques to make informed decisions. By considering multiple criteria, such as road condition, traffic volume, and budget constraints, decision-makers can objectively prioritize maintenance projects and allocate resources effectively.
f.
Objective evaluation: MCDM techniques provide a systematic and objective evaluation framework for road maintenance projects. By considering various criteria and assigning weights to each criterion, decision-makers can ensure transparency and fairness in the decision-making process, minimizing bias and subjective judgments.
g.
Long-term planning: image processing combined with MCDM facilitates long-term planning for road maintenance. Historical data and condition assessments can be analyzed to identify maintenance patterns, predict future maintenance needs, and develop proactive maintenance strategies, leading to improved road network performance and reduced overall lifecycle costs.
Conclusion
The prioritization of pavement distress is crucial for transportation agencies to ensure the safety and comfort of road users. This paper introduces a novel approach that integrates image processing techniques and the Analytic Hierarchy Process (AHP) to effectively identify and prioritize different types of pavement distress. By capturing high-resolution images of the pavement surface and analyzing them using image processing algorithms, the severity of cracks, potholes, and rutting can be quantified. The AHP is then utilized to prioritize this distress based on its severity and impact on road users. The results of testing on real-world pavement images demonstrate the effectiveness of the proposed approach in accurately identifying and prioritizing pavement distress. In addition, several scenarios were assumed such as:
(1)
the imposition of a fixed time constraint.
(2)
the imposition of a fixed budget constraint.
(3)
the absence of any fixed time or budget constraints to demonstrate the proposed method's capabilities.
In the first case, pavement maintenance works were completed by 37.5% of the specific sector in 12 days and $487.3 cost.
In the second case, pavement maintenance works were completed by 62.5% of the specific sector in 18 days and $947.6 cost.
In the third case, pavement maintenance works were completed by 100% of the specific sector in 24 days and $1532 cost.
Automation may eventually be more cost-effective than human detection techniques, especially for large-scale projects or for organizations in charge of maintaining a large number of roadways, even if they may need an initial investment in hardware and software. Automated pavement crack detection has several benefits over human detection techniques, including speed, consistency, accuracy, data management, and cost-efficiency. Automated systems are a useful tool for evaluating pavement conditions, prioritizing maintenance requirements, and organizing upcoming repairs, that has high accuracy and save time and cost compared to the manual method. These novel and unique capabilities provide much-needed support for transportation agencies in decision-making to create informed decisions about maintenance and repair activities, leading to improved road safety, reduced maintenance costs, and ensuring the longevity and quality of road infrastructure.
Declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Ethical approval
This article doesnot contain any studies with human participants or animal performed by any of the authors.
Informed consent
For this type of study formal consent is not required.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
