Pavement Inspection Research Articles

An image-based system for asphalt pavement bleeding inspection

ABSTRACT The pavement distress information is essential for evaluating the pavement. Automatic pavement inspection systems can be so beneficial in providing needed information on pavement conditions. The bleeding is one of the asphalt pavement distresses and directly affects the pavement skid resistance and road safety. This study aims to develop an image-based system for the comprehensive evaluation of pavement bleeding. This evaluation consists of three main parts: bleeding occurrence detection, bleeding region segmentation, and severity-based classification of bleeding regions. For implementing the proposed system, deep learning-based model and transfer learning method are used in the detection part, and wavelet transform is the main process in the segmentation part. Also, gray-level co-occurrence matrix, wavelet transform, and signal to noise ratio are used to extract image texture-based features from various levels of the bleeding severity (low, mid, and high). Then a decision tree has been made based on extracted features for severity-based classification. The proposed system can indicate the bleeding index based on distress density and severity. Results show good performance in detection, segmentation, and severity-based classification parts, on the average of performance indices with 98%, 89%, and 93%, respectively. Therefore, the proposed system provides an efficient method for comprehensive pavement bleeding evaluation.

Study of wavelet entropy for airport pavement inspection using a multistatic ground-penetrating radar system

The multilayer nature of airport pavement structures is susceptible to the generation of voids at the bonding parts of the structure, which is also called interlayer debonding. Observations have shown that the thickness of the resulting voids is usually at the scale of millimeters, which makes it difficult to inspect. The efficient and accurate characteristics of ground penetrating radar (GPR) make it suitable for large area inspections of airport pavement. In this study, a multistatic GPR system was used to inspect the interlayer debonding of a large area of an airport pavement. A special antenna arrangement can obtain common-midpoint (CMP) gathers during a common-offset survey. The presence of interlayer debonding affects the phase of the reflection signals, and the phase disturbance can be quantified by the wavelet transform. Therefore, we have developed an advanced approach that uses the average entropy of the wavelet transform parameters in CMP gathers to detect the interlayer debonding of airport pavement. The results demonstrate that the regions with high entropy correspond to regions where tiny voids exist. The new approach introduced in this study was then evaluated by a field-base experiment at an airport taxiway model. The results show that our approach can detect interlayer debonding of the pavement model accurately and efficiently. On-site coring results confirm the performance of our approach.

Flexible Pavement Crack’s Severity Identification and Classification using Deep Convolution Neural Network

Effective road maintenance program is vital to ensure traffic safety, serviceability, and prolong the life span of the road. Maintenance will be carried out on pavements when signs of degradation begin to appear and delays may also lead to increased maintenance costs in the future, when more severe changes may be required. In Malaysia, manual visual observation is practiced in the inspection of distressed pavements. Nonetheless, this method of inspection is ineffective as it is more laborious, time consuming and poses safety hazard. This study focuses in utilizing an Artificial Intelligence (AI) method to automatically classify pavement crack severity. Field data collection was conducted to allow meaningful verification of accuracy and reliability of the crack’s severity prediction based on AI. Several important phases are required in research methodology processes including data collection, image labelling, image resizing, image enhancement, deep convolution neural network (DCNN) training and performance evaluation. Throughout the analysis of image processing results, the image output was successfully classified using MATLAB software. The good agreement between field measurement data and DCNN prediction of crack’s severity proved the reliability of the system. In conclusion, the established method can classify the crack’s severity based on the JKR guideline of visual assessment.

Deep Learning Based Pavement Inspection Using Self-Reconfigurable Robot.

The pavement inspection task, which mainly includes crack and garbage detection, is essential and carried out frequently. The human-based or dedicated system approach for inspection can be easily carried out by integrating with the pavement sweeping machines. This work proposes a deep learning-based pavement inspection framework for self-reconfigurable robot named Panthera. Semantic segmentation framework SegNet was adopted to segment the pavement region from other objects. Deep Convolutional Neural Network (DCNN) based object detection is used to detect and localize pavement defects and garbage. Furthermore, Mobile Mapping System (MMS) was adopted for the geotagging of the defects. The proposed system was implemented and tested with the Panthera robot having NVIDIA GPU cards. The experimental results showed that the proposed technique identifies the pavement defects and litters or garbage detection with high accuracy. The experimental results on the crack and garbage detection are presented. It is found that the proposed technique is suitable for deployment in real-time for garbage detection and, eventually, sweeping or cleaning tasks.

Design of airport road surface inspection system based on machine vision and deep learning

The specific objective of this study was to solve the problem of high risk and low efficiency of manual inspection of airport road surface. This paper describes the design of an Airport Road Surface Intelligent Inspection System Based on machine vision and deep learning. The system completes the airport pavement inspection through autonomous navigation. MobileNet-SSD and Mask R-CNN algorithms are used for target detection and semantic segmentation of airport pavement apparent disease and foreign object bebris in this paper. At the same time, topological feature extraction and pixel size measurement were carried out. Finally, the information is uploaded to the storage server and the target information is displayed on the Android mobile terminal. The inspection system can complete the work of airport pavement inspection and provide a new way for airport pavement health monitoring.

Pavement Distress and Debris Detection using a Mobile Mapping System with 2D Profiler LiDAR

Regular pavement monitoring over highways and airport runways is vital for public agencies to ensure the safe riding of vehicles and aircrafts. Highways are mostly subject to cracking and potholes along with a few instances of debris around construction work zones. Airports are also concerned with debris but have much lower tolerance for the presence of foreign object debris (FOD) that could possibly damage the aircraft. LiDAR is rapidly emerging in a variety of mobile mapping systems (MMS) and will likely be integrated into many transportation vehicles over the next decade for pavement inspection. This paper proposes a unique algorithm for pavement surface inspection with the help of MMS driven at highway speeds. The study analyzed LiDAR data acquired for 8 mi of highway collected at approximately 55 to 60 mph. This study indicates that an adequately designed MMS along with the proposed algorithm can efficiently detect pavement anomalies as small as 2 cm in the form of cracking, potholes, surface debris, or any combination of these. This is more than sufficient for highways, where debris such as ladders and tires are an order of magnitude larger. For evaluating the effectiveness of detecting smaller airport FOD, a validation dataset was created by driving the MMS at 15 mph adjacent to a debris field of 50 sample pieces of FOD collected from an airport. The study found that 100% of the FOD items larger than 2 cm in size (12 out of 50 samples) were detected successfully at 15 mph. Both datasets suggest that MMS LiDAR is sufficient for pavement inspection and as sensor fidelity increases, even small FOD will be able to be detected with the algorithm proposed in this paper.

Combination of pixel-wise and region-based deep learning for pavement inspection and segmentation

ABSTRACT This paper presents a new model combining pixel-wise and region-based deep learning to provide a pavement inspection technology for jointly obtaining the distress classes, locations, and geometric information. This proposed model, called the segmentation RCNN, added an extract branch in a faster region convolutional neural network (Faster RCNN) for assigning the pixels in each region of interest (ROI) from an image into one of the pavement distresses or background, in parallel with the existing branches for ROI classification and bounding-box regression in the Faster RCNN. The effectiveness of the proposed model was tested by a pavement-image database collected from 16 asphalt pavements. The results indicated that the proposed model detected and segmented the pavement distresses (cracks, potholes, and patches) with mean intersection over unions of 87.6% and 70.3%, respectively. The proposed model was stable under various real-world conditions. The model reduced the computation costs, which provided a novel direction to achieve real-time pavement inspection.

Field Experiment for Accuracy Verification of the Devices Measuring International Roughness Index in TRUE Project: Five Years Report

A lot of devices measuring the International Roughness Index (IRI) has recently been developed and applied in response to the pavement inspection purposes. Consequently, the accuracy of measured IRIs obtained with different devices on the same road surface has become a practical concern for providing a consistent inspection result. Against this background, the specified non-profit organization Pavement Diagnosis Researchers Group (PDRG) in Japan has undertaken a project regarding the experiment to harmonize and compare Test methods for surface Roughness Under actual road Environment (TRUE project) since 2014. The TRUE project has so far conducted the roughness measuring experiments three times over five years at Hokkaido, Japan. This paper introduces the features of the TRUE project such as participated devices, test site selection, and data recording and reporting procedures. A benchmark testing for obtaining reference profiles is also provided to make the project reliable. In this paper, the results of experiments conducted in the last five years are statistically described as well. Finally, this paper introduces the potential of the TRUE project such as a new device grouping and the accuracy evaluation report for device certification for contributing to further improvement of a device participated in the project.

Automatic Detection of Cracks in Asphalt Pavement Using Deep Learning to Overcome Weaknesses in Images and GIS Visualization

The crack ratio is one of the indices used to quantitatively evaluate the soundness of asphalt pavement. However, since the inspection of pavement requires much labor and cost, automatic inspection of pavement damage by image analysis is required in order to reduce the burden of such work. In this study, a system was constructed that automatically detects and evaluates cracks from images of pavement using a convolutional neural network, a kind of deep learning. The most novel aspect of this study is that the accuracy was recursively improved through retraining the convolutional neural network (CNN) by collecting images which had previously been incorrectly analyzed. Then, study and implementation were conducted of a system for plotting the results in a GIS. In addition, an experiment was carried out applying this system to images actually taken from an MMS (mobile mapping system), and this confirmed that the system had high crack evaluation performance.

An image-based system for pavement crack evaluation using transfer learning and wavelet transform

Automatic systems for pavement inspection can significantly enhance the performance of the Pavement Management Systems (PMSs). Cracking is the most current distress in any type of pavement. Progress of various technologies leads to a lot of effort in developing an automatic system for pavement cracking inspection. In the early image-based systems, the feature extraction process for crack classification must be done by using various image processing techniques in an expert-based system. In recent years, the new machine learning techniques such as a deep convolutional neural network (DCNN) provide more efficient models with the ability of automatic feature extracting, but these models need a lot of labeled data for training. Transfer learning is a technique that solves this problem using pre-trained models. In this research, several pre-trained models (AlexNet, GoogleNet, SqueezNet, ResNet-18, ResNet-50, ResNet-101, DenseNet-201, and Inception-v3) have been used to retrain based on pavement images using transfer learning. This study aims to evaluate the efficiency of retrained DCNNs in the detection and classification of the pavement cracking. Also, it presents a more effective algorithm based on a developed wavelet transform module with more regulizer parameters for crack segmentation. The result indicated that retrained classifier models provide reliable outputs with a range of 0.94 to 0.99 in confusion matrix-based performance, but the speed of some models is significantly higher than others. Also, the results clarified that the developed wavelet module could segment crack pixels with a high level of clarity.

Ultrasonic inspection of asphalt pavements to assess longitudinal joints

Longitudinal joints existing in between the lanes of asphalt pavements may initiate deterioration. Traditionally, core density, deflection, and nuclear density tests are used for the quality control. However, such techniques may not suit to the surface at the joints to assess their condition. Alternatively, the ultrasonic surface wave (USW) methods have the potential to both assess the longitudinal joints and estimate the pavement thickness at the same time. In this study, the USW are investigated on two lab-scale asphalt slabs (one laboratory prepared, and the other is cut from an as-built pavement) and on an in-service asphalt pavement to develop an ultrasound-based assessment methodology. Initially, an empirical compaction model is developed to produce the custom-size slab with the desired air-void profile to mimic a pavement with joint. Then, a variety of coupling systems between the pavement and the ultrasonic transducers are trialed to determine the optimum one. The recorded data are processed to determine the dispersion in velocity and the attenuation, which are then interpreted to estimate the pavement thickness and assess the joint quality, respectively. The dispersion curve is found capable of determining the pavement thickness with a precision of 1 cm, while the attenuation curve is observed to be affected by the transducer configuration excessively. Therefore, a normalisation technique, named the Fourier transmission coefficient (FTC), is implemented to reduce the undesired variability caused by the transducer coupling and type. Finally, it is demonstrated on an as-built pavement that the FTC has promising potential to detect, and hence evaluate the quality of longitudinal joints.

A novel pavement mean texture depth evaluation strategy based on three-dimensional pavement data filtered by a new filtering approach

Automatic pavement inspection is a reliable strategy to evaluate pavement conditions. Generally, the skid resistance of asphalt pavement is employed to characterize the pavement surface characteristics. The mean texture depth (MTD) of pavement is a crucial metric to characterize the macrotexture of asphalt pavement. The primary aim of this article is to evaluate pavement MTD. To evaluate this parameter based on three-dimensional (3-D) point cloud data extracted from a self-developed 3-D detection system, a novel filtering method and reference surface integral method were proposed. The novel filtering method was first employed to eliminate the isolated noise of the extracted 3-D point cloud data and was compared with the primary surface profile (PSP)-based filtering method. Then, a 3-D virtual model of pavement texture was generated based on the filtered 3-D data. Finally, a novel method, the reference surface integral, was proposed to evaluate the pavement MTD. The evaluated results were compared with the results measured via manual sand-patch testing. The results show that the novel filtering method performs better filtering performance than PSP-based filtering. The error between the evaluated MTD and the measured MTD is within 3.28%, which demonstrates that the proposed filtering method and MTD evaluation method have robust advantages and can be employed to evaluate pavement performance.

RUNWAY PAVEMENT INSPECTIONS USING DRONE - SAFETY ISSUES AND ASSOCIATED RISKS

RUNWAY PAVEMENT INSPECTIONS USING DRONE - SAFETY ISSUES AND ASSOCIATED RISKS

Development and application of a non-destructive pavement testing system based on linear structured light three-dimensional measurement

Automatic detection dramatically improves the efficiency and accuracy of asphalt pavement inspection. However, the existing detection systems cannot coordinate the balance between efficiency and accuracy. Three-dimensional pavement detection was proposed and developed in this article. The theoretical model of camera imaging was first established based on optical triangulation. The camera’s intrinsic and extrinsic parameters were evaluated by utilizing a checkerboard calibration target, and the camera lens distortion was also considered to ensure the accuracy of the three-dimensional measurements. The subpixel coordinates of the single-stripe center were extracted based on two-dimensional image processing technology, and then the pixel coordinates were converted into real-world elevation coordinates based on the camera imaging model. The cubic spline method and the three-dimensional data filtering method based on the pavement surface profile were applied to restore the data and remove the noise, respectively. Then the three-dimensional reconstruction was achieved based on Delaunay triangulation algorithm. We expect that this new system can significantly reduce the costs without decreasing the test accuracy, thus making large-scale engineering applications viable.

Analysis of Optimal Flight Parameters of Unmanned Aerial Vehicles (UAVs) for Detecting Potholes in Pavements

Pavement maintenance seeks to provide optimal service conditions. Before maintenance, it is necessary to know the condition of the pavement by inspection, a crucial step in deciding on the repair to be carried out. In this sense, unmanned aerial vehicles (UAVs) seem to be an economic substitute compared to the ground laser scanner for pavement inspection tasks. This research seeks to develop a method to measure potholes using 3D models generated with photographs acquired by a UAV and process them using a software based on the Structure from Motion-MultiView Stereo (SfM–MVS) technique. The contribution of this document is the proposal of recommendations for the acquisition of photographs for the realization of the models. To develop these recommendations, an experiment was carried out to evaluate the accuracy in the reconstruction of 3D models using images obtained from the variation and combination of flight planning parameters and data capture. Then, to validate these recommendations, a bumpy section of pavement was modeled using the SfM–MVS method. The results show that for heights of 10 and 15 m the use of this methodology is applicable for the measurement of the width and depth of potholes.

Detection of Pothole by Image Processing Using UAV

Potholes on road can generate costly damage to flat tire or can cause wheel damage of motor cycle. Especially in India vehicle collision and major accidents happens due to immense depth of pothole on road. Thus, detecting and repairing of potholes on road is major challenge for in ITS (Intelligent Transportation System) service and road management system. In order to enhance the efficiency of pavement inspection,currently some new sorts of remote sensing methods without any physical damage to the pavements is possible. In our study the pavement images captured by the unmanned aerial vehicle are collected and distinguish the pavements with the pothole from the normal pavements by performing the Morphological Operations on the images. UAV is new tool under the Remote sensing technique for the pavement inspection with more efficiency as it is usable in hard to reach places also where road maintenance is difficult. Through large transport surveillance, we built a method that can detect pothole using various algorithms of image processing for time optimization of surveillance. A model of potholes is constructed using the image library, which is used in an algorithmic approach that combines a live road footage and simultaneously detection of potholes using simple image processing techniques such as a medium filter and edge detection. Using this approach, it was possible to detect potholes with a precision of 80% and recall of 74.4.%. Using UAV, we can reach in remote places and specially to National Highways, detection of pothole and collecting all database become very easy and time saving. With more advanced technologies such as Artificial Intelligence and Machine Learning algorithms can be used for surveillance of road with more precise results with automatic repairing robots or machines

Proposal of an Airport Pavement Maintenance Management System for Cape Verde

In order to ensure safe and efficient operation of aircrafts, airport operators are faced with the need to maintain airport runway, taxiway and apron pavements in good condition, within a given budget. In this context, this paper describes the first developments and the methodological approach of a doctoral work with the following objectives: to improve and validate in-vehicle technique for pavement inspection using GNSS, video image capture and Geographic Information System, increasing the degree of automation of pavement distress identification; to develop a procedure for the definition of preventive maintenance strategies for airport pavements, adapted to Cape Verde airport system, using modelling techniques to predict pavement performance and definition of service and trigger levels; and also to develop or calibrate an existing computer application that can be used as a decision support tool.

Classification of pavement crack types based on square bounding box diagonal matching method

With the increase in traffic volume, it brings new problems to people’s live, that is, the task of inspection and maintenance of road pavements has become increasingly arduous. Therefore, the detection and identification of the road surface has become particularly urgent. This paper proposes a crack classification method based on diagonal matching of square bounding boxes. This method is used to match the geometric features of different types of cracks and gives an evaluation criterion for the regional similarity metrics that can be used to construct the classification of crack types. It provides quantitative criteria for refining the intrinsic relationship of complex types of crack areas. The experimental results show that the classification function of the crack type of the pavement is better. The results show that the accuracy of the identification and classification of block and crack cracks is obviously better than the traditional classification methods such as random forest and support vector machine.

An integrated approach to automatic pixel-level crack detection and quantification of asphalt pavement

Discovering and assessing cracks is widely thought to be critical for maintaining the healthy conditions of asphalt pavement. Unfortunately, the inspection of pavement for cracks is not only labor-intensive, time-consuming, inefficient, and costly, but it is also unable to detect and quantify cracks accurately at the pixel level. To address this problem, we propose an integrated approach based on the convolutional neural network DeepLabv3+ for crack detection, as well as a crack quantification algorithm for crack quantification at the pixel level. The quantification algorithm is used to evaluate five important indicators: crack length, mean width, maximum width, area, and ratio. To fully verify the performance of DeepLabv3+, 50 images were studied; the best image showed a mean intersection of union (MIoU) of 0.8342. For testing, 80 new images (including both asphalt pavement images and concrete pavement images) were used. DeepLabv3+ was found to be reliable and widely applicable for crack detection, and it demonstrated an MIoU of 0.7331. Of the various quantitative indicators, the crack length had the lowest relative error rate of the predicted values and therefore had the highest accuracy (its relative error rate ranged from −25.93% to 14.11%). We also compared our system with four state-of-the-art methods. The results showed our integrated approach to be more effective and more accurate in both the detection and quantification of cracks. The integrated approach could potentially serve as the basis of an automated, cost-effective pavement-condition assessment scheme for the operation and maintenance of pavement.

Application of georadars for detecting subsurface defects in layers of non-rigid road pavements

The relevance of the problem under consideration is a consequence of the high cost of classical methods of pavement inspection. At the same time, the use of modern pulse georadars allows to ensure a relatively low cost of monitoring the current state of highways, since it allows to obtain georadar data when a laboratory vehicle is moving at the speed of a traffic flow. This minimizes the role of costly and time-consuming operations such as coring or cutting. The purpose of this work is to improve the methods for obtaining primary GPR data, which were previously proposed by the authors to improve the accuracy and reliability of the results of processing pulsed GPR signals. Materials and methods. When processing model and experimental data, first of all, modern theoretical methods of processing pulse signals from ground penetrating radars, as well as methods of computer modeling, were used. Results. Based on the analysis of the factors that determine the key features of GPR signals, a signal calibration method has been proposed, which makes it possible to increase the reliability of detecting such defects in layers of non-rigid road pavement made of monolithic materials, such as loss of interlayer adhesion, or identification of thin layers from an electrophysical point of view in multilayer media. Conclusions. Combining the signal calibration method together with the previously proposed approach to detecting the loss of interlayer adhesion and the performed numerical simulation made it possible to increase the reliability of the procedure for non-destructive testing of road pavements and other building structures. During the work, laboratory experiments were performed on model structures. The analysis of the obtained data was performed using the developed software GeoVizy.