Paramarthalingam, Arjun

Paramarthalingam A., Janarthanan A., Arivunambi A., Ariyangavu S.S., Senthamaraikannan H., Ganapathy R.

This article introduces an Artificial Intelligence (AI) enabled virtual mouse system that utilizes hand gestures and fingertip detection to operate computer mouse functions through AI and computer vision techniques. It serves as a convenient alternative to traditional physical mouse, offering users increased flexibility and accessibility in navigating and controlling their devices. In this article, the proposed work uses three modules such as OpenCV, MediaPipe and PyautoGUI to create the virtual mouse system. OpenCV library is used for its real-time computer vision functionality to help us to capture the hand using the web camera. The MediaPipe framework is used to detect the hand region using KLT Tracking Algorithm and then the K-cosine border tracking algorithm is used to identify various types of hand gesture movements to mimic the computer mouse cursor movement and scrolling operations. The PyAutoGUI module is used to perform appropriate mouse actions based on the recognized hand gestures.

Paramarthalingam A., Sivaraman J., Theerthagiri P., Vijayakumar B., Baskaran V.

Visually impaired individuals encounter numerous impediments when traveling, such as navigating unfamiliar routes, accessing information, and transportation, which can limit their mobility and restrict their access to opportunities. However, assistive technologies and infrastructure solutions such as tactile paving, audio cues, voice announcements, and smartphone applications have been developed to mitigate these challenges. Visually impaired individuals also face difficulties when encountering potholes while traveling. Potholes can pose a significant safety hazard, as they can cause individuals to trip and fall, potentially leading to injury. For visually impaired individuals, identifying and avoiding potholes can be particularly challenging. The solutions ensure that all individuals can travel safely and independently, regardless of their visual abilities. An innovative approach that leverages the You Only Look Once (YOLO) algorithm to detect potholes and provide auditory or haptic feedback to visually impaired individuals has been proposed in this paper. The dataset of pothole images was trained and integrated into an application for detecting potholes in real-time image data using a camera. The app provides feedback to the user, allowing them to navigate potholes and increasing their mobility and safety. This approach highlights the potential of YOLO for pothole detection and provides a valuable tool for visually impaired individuals. According to the testing, the model achieved 82.7% image accuracy and 30 Frames Per Second (FPS) accuracy in live video. The model is trained to detect potholes close to the user, but it may be hard to detect potholes far away from the user. The current model is only trained to detect potholes, but visually impaired people face other challenges. The proposed technology is a portable option for visually impaired people.

Paramarthalingam A., Arivunambi A., Janarthanan A.

The ever-growing expansion of highway transportation infrastructure leads to significant role in development of every nation. The number of people owning four-wheeler vehicle has increased every year with automatic driving cars attracts people, further it will make a great impact in short to long distance travel. The autonomous vehicles generally use lane and curve detection approaches to automatically discover the way using deep learning algorithms. To tackle the problem of inadequate robustness for extracting multi-lane marking, a multi-lane identification strategy relying on a deep neural network based on FCN network is proposed in this paper. The ResNet-152 based fully convolutional network is used for lane and curve detection. The ResNet-152 model is chosen in this work because the existing models viz. VGG-16, AlexNet performs worse when number of layers increased, further error rate also get increased, but in ResNet-152 error rate decreases with increasing layers. The ResNet-152 model used skip connections. Finally, the fit interval is determined using the Hough Transform, and the lane marker is fitted using the least square method. The experiments are carried out on the TuSimple dataset. The suggested FCN-based lane and curve detection approach performs well on the detection of curved lane roads.

Banu M. S., S. R. S., Duraikannu G., Paramarthalingam A.

Frequency of Itemsets plays a crucial role in analytics of retail industry, which delves into latent patterns in customer purchasing behavior. This paper presents an Apriori algorithm to extract associations among products in a given dataset, shedding light on frequently co-occurring items. By discerning these relationships, the business gains profound insights into customer preferences and tendencies, aiming not only to understand current purchasing behavior but also to identify potential cross-selling opportunities. As businesses rely on transactional data for insights, analysis reliability hinges on data quality. This study explores missing values, outliers, and data inconsistency, impacting market basket analysis accuracy. Leveraging the Apriori algorithm facilitates the revelation of robust product associations, enabling strategic optimization and heightened customer satisfaction. The gleaned insights inform targeted marketing, product placements, and inventory management, catalyzing more effective business optimization in the retail sector.

Venkatakrishnan R., Sivagurunathan M., Siva L., Subramani S., Paramarthalingam A.

This study presents a comprehensive analysis of company registration trends in India, focusing on data sourced from the Registrar of Companies (RoC). The study investigates the temporal patterns of company registrations, emphasizing principal business activities and classifications into public and private sectors. Leveraging advanced data pre-processing techniques, The Study explores the spatial and temporal dynamics of company registrations. Furthermore, a suite of machine learning algorithms, including Linear regression, Decision tree, Random Forest, GBM, SVM, KNN, Naive Bayes, and Weka, is employed to predict future registration trends. Visualization of insights is facilitated through the use of Tableau. The findings provide valuable insights for stakeholders in business, policy, and investment sectors, aiding informed decision-making and strategic planning.

R R.K., Paramarthalingam A.

Arivunambi A., Paramarthalingam A.

Wireless Sensor Networks (WSNs) depicts to the group of spatially distributed sensor nodes, each node is having individual battery for power consumption which senses the factors like Temperature, Pressure, Soil makeup, Vehicular movement, and checks the physical presence of objects, which are communicate through internet and passing messages between nearest neighbor nodes. WSN are extensively used in several applications like Environmental Monitoring, Area Monitoring, Engineering Monitoring, Health Care Monitoring, and Military Applications. Since WSN are highly susceptible for many attacks especially Jamming Attack. This causes high energy consumption, reduces the life time of nodes, affects the entire sensor networks and most significantly also the confidentiality and integrity of the data packets which are transmitted through dissimilar nodes in the network. In order to overcome these issues, a new bio-inspired algorithm called Intelligent Slime Mold (ISM) has been introduced for detection jamming node and to setting up of alternate optimal path form source node to destination node in WSN to avoid jammer affected nodes by applying Context Aware Decision Rule (CADR). This work ensures the confidentiality and integrity of both the network and data packets.

Janarthanan A., Paramarthalingam A., Arivunambi A., Vincent P.M.

Arivunambi A., Paramarthalingam A.

Current trending- Internet of things (IoT) is internetworking of an assortment of hardware devices to offer a collection of applications and services. In the present-day world, ransomware cyber-attack has become one of the major attacks in IoT systems. Ransomware is a hazardous malware that targets the user’s computer inaccessible or inoperative, and then requesting the computer victim user to transfer a huge ransom to relapse the damage. At instance, the evolution rate outcomes illustrate that the level of attacks such as Locky and Cryptowall ransomware are conspicuously growing then other ransomware. Thus, these ransomware relations are the latent threat to IoT. To address the issue, this paper presents Two-phase ransomware prediction model based on the behavioral and communication study of Cryptowall ransomware for IoT networks. This proposed Two-phase model equipped with, Phase-1: observes the inward TCP/IP flowing traffic through a monitoring server to avert the ransomware attack The procedure of the monitoring server is to monitor the IoT's TCP/IP. The process of Monitoring TCP/IP is to extract TCP/IP header and routines command and control (C&C) server IP blacklisting to discover the ransomware attacks. In Phase-2: the proposed system will also analyze the application pattern for malicious behavior of the Web and URLs. Several societies have very affluent security tools in their milieu, but their events or logs are not monitored, which make affluent tools ineffective. The process of having efficient security based monitoring server is vital for detecting and controlling the ransomware attack.

Paramarthalingam A., Arivunambi A., Thapasimony S.

Presently, the urban world are switching to rooftop farming with technology support to cope with the increasing food demand and effective utilization of various resources. But the monitoring of farming in high raised building (rooftop) is little challenging. This paper presents an IoT based smart rooftop irrigation system to efficiently manage water dispersal and provides improved urban based rooftop farming productivity. The proposed model regularly monitors soil temperature and moisture level, i.e., rooftop farming are irrigated automatically without physical presence of planter and it also uses smart mobiles for irrigation control. The proposed automatic irrigation farming modal uses sensor technology, communication technology and embedded hardware technology. That is, the atmospheric weather data such as moisture level and water level are collected periodically from different part of rooftop farm area and it is analyzed, which will trigger the switching of water motor/overhead tank pump and send status updates to the planter.

Arivunambi A., Paramarthalingam A., Sanju P., Uthayashangar S., L K.V.

Virtual mind inquiries about hastening the improvement of reasonable incessant Brain Computer Interface (BCI). Equipment developments the expansion ability of Virtual mind dissect and Brain PC-wearable sensors have made possible a few novel programming systems for engineers to use and make applications joining BCI and IoT(Internet of Things). At present, a comprehensive study on BCI in IoT from dissimilar viewpoints; together with Electroencephalogram (EEG) based BCI models, and current dynamic stages. In view of this analyses, the fundamental discoveries of study eye on three substantial improvement patterns of BCI, which are EEG, IoT, and cloud computing. Utilizing this it is totally helpful for finding the genuine condition regardless of whether the cerebrum is alive or dead. In the incident that it is active, at that point the movement of the mind is checked and put away. Through this anybody can arrive at resolution that whether the activity done is legitimate or illegal. This has a favorable position for two situations. First is for Adulteration in bank subtleties and secondly fabrication in resource archives. The principle point is to transfer human cerebrum private things in the cloud, if there are any adjustments in the status of the mind, the virtual cerebrum site will go about as the human mind.

Paramarthalingam A., Thankanadar M.

Arjun P., Stephenraj S., Kumar N.N., Kumar K.N.

Today's modern world people preferred to live the sophisticated life with all facilities. The science and technological developments are growing rapidly to meet the above requirements. With advanced innovations, Internet of Things (IoT) plays a major role to automate different areas like health monitoring, traffic management, agricultural irrigation, street lights, class rooms, etc., Currently we use manual system to operate the street lights, this leads to the enormous energy waste in all over the world and it should be changed. In this survey we studied about, how IoT is used to develop the street lights in the smart way for our modern era. It is an important fact to solve the energy crises and also to develop the street lights to the entire world. In addition, with the study on smart street lighting systems we analyzed and described different sensors and components which are used inIoT environment. All the components of this survey are frequently used and very modest but effective to make the unswerving intelligence systems.

Arjun P., Mirnalinee T.T.

This paper describes a multi-scale feature integration framework using angular pattern (AP), binary AP (BAP) and sequential backward selection (SBS) algorithms. These angular descriptors are represented by multi-scale features from which the best subsets of the scales are chosen using five-fold cross-validation technique along with SBS algorithm for efficient image retrieval. The SBS algorithm reduces the dimensionality of feature space which in turn reduces the matching time complexity. The extracted AP and BAP features are represented in histograms and are compared by the Chi-square distance metric. The experimental analysis is performed on the MPEG-7 CE-1 Part-B dataset images to demonstrate the effectiveness of multi-scale feature integration using SBS algorithm. The image retrieval performance of this framework is compared with state-of-the-art shape descriptors. Being multi-scale global shape descriptors, the proposed framework captures complete information about the shape and are invariant to scaling and rotation transformations.

Park S., Nguyen N.

Bodake V.J., Meeeshala S.K.

Elavarasi T., Nagarajan R.

Paramarthalingam A., Janarthanan A., Arivunambi A., Ariyangavu S.S., Senthamaraikannan H., Ganapathy R.

This article introduces an Artificial Intelligence (AI) enabled virtual mouse system that utilizes hand gestures and fingertip detection to operate computer mouse functions through AI and computer vision techniques. It serves as a convenient alternative to traditional physical mouse, offering users increased flexibility and accessibility in navigating and controlling their devices. In this article, the proposed work uses three modules such as OpenCV, MediaPipe and PyautoGUI to create the virtual mouse system. OpenCV library is used for its real-time computer vision functionality to help us to capture the hand using the web camera. The MediaPipe framework is used to detect the hand region using KLT Tracking Algorithm and then the K-cosine border tracking algorithm is used to identify various types of hand gesture movements to mimic the computer mouse cursor movement and scrolling operations. The PyAutoGUI module is used to perform appropriate mouse actions based on the recognized hand gestures.

R R.K., Mathew A.M.

Venkatraman A., Ruchitha A., Sellamuthu S., Shalini D., Reddy P.D.

Witczak D., Szymoniak S.

Karaklas N., Papavasileiou A., Makris S.

McIntosh T., Susnjak T., Liu T., Xu D., Watters P., Liu D., Hao Y., Ng A., Halgamuge M.

Ransomware has grown to be a dominant cybersecurity threat by exfiltrating, encrypting, or destroying valuable user data and causing numerous disruptions to victims. The severity of the ransomware endemic has generated research interest from both the academia and the industry. However, many studies held stereotypical assumptions about ransomware, used unverified, outdated, and limited self-collected ransomware samples, and did not consider government strategies, industry guidelines, or cyber intelligence. We observed that ransomware no longer exists simply as an executable file or limits to encrypting files (data loss); data exfiltration (data breach) is the new norm, espionage is an emerging theme, and the industry is shifting focus from technical advancements to cyber governance and resilience. We created a ransomware innovation adoption curve, critically evaluated 212 academic studies published during 2020 and 2023, and cross-verified them against various government strategies, industry reports, and cyber intelligence on ransomware. We concluded that many studies were becoming irrelevant to the contemporary ransomware reality and called for the redirection of ransomware research to align with the continuous ransomware evolution in the industry. We proposed to address data exfiltration as priority over data encryption, to consider ransomware in a business-practical manner, and recommended research collaboration with the industry.

Witczak D., Szymoniak S.

The Internet of Things is currently one of the fastest-growing branches of computer science. The development of 5G wireless networks and modern data transmission protocols offers excellent opportunities for rapid development in this field. The article presents an overview of monitoring and control systems based on the Internet of Things. The authors discuss various aspects of these systems, including their architecture, applications, and challenges. We focus on analyzing the latest achievements in this field, considering technological innovations and practical applications in various sectors. Also, we emphasize the importance of integrating data from multiple sources and developing data analysis algorithms to ensure the effectiveness and precision of IoT-based monitoring and control systems. The article provides a valuable overview of the current state of knowledge in this dynamic area, inspiring further research and technological development. It also includes case studies showing various IoT device applications and energy consumption management.

Bhisikar S., Jagtap P., Sonawane S., Sawant H., Pisal P.

Asbai M., Ghilane H., Alaoui L.L.

Smart lighting is a gateway to the smart cities of the future. Smart lighting and IoT are the keys to the city of the future. By using these technologies, cities can improve safety, reduce costs, and increase energy efficiency. However, smart cities require a lot of work. New technologies are expensive, local governments are constrained, and politics is geared toward short election cycles, making it difficult to establish a highly operational and financially efficient centralized technology deployment model that is reused in areas urban areas on a global or national scale. In Morocco, we note an awareness regarding the installation of intelligent street lighting in different areas of the kingdom and the development of smart cities. However, there are questions about the return on investment of these smart lighting solutions. This work seeks to evaluate, through a qualitative study, the ROI of the implementation of intelligent public lighting in certain areas of the kingdom to identify the advantages and challenges encountered. The main results claim that smart lighting allows real-time energy monitoring, automatic adjustment of brightness according to weather conditions and human activity, and the specific need at the location of the light point.

Wijaya C., Andriyadi A., Chen S., Wang I., Yang C.

With a growing emphasis on environmental health and safety, monitoring and managing air quality in large-scale settings such as campuses are becoming increasingly critical. This research proposes an innovative approach that integrates multilocation Internet of Things (IoT) sensors, edge artificial intelligence (AI), machine learning, and public API integration to create a comprehensive air quality monitoring and notification system for campus environments. Our framework deploys IoT sensors across various locations within the campus to collect real-time data on air quality parameters. Leveraging edge AI capabilities, these sensors process data locally, enabling rapid analysis and anomaly detection without the need for centralized processing. Furthermore, machine learning algorithm is used to analyze the collected data, identify patterns, and predict air quality. To enhance user accessibility and engagement, the system use public APIs to deliver notifications and alerts regarding air quality status.

Sivakarthi G., Gobinath A., Abikannan P.R., Balasubramani T., Baranisri K., Earlene Melba J., Harinisri T.R., SanthoshKumar V., Manjula Devi C.

Lu Y., Yi K., Xu Y.

Fruit fly species classification is a fine-grained task as there is a small gap between species. In order to effectively identify and improve the recognition of fruit flies, a fine-grained image-recognition method based on a multi-channel self-attention mechanism was studied and a network framework for fine-grained image recognition based on deep learning was designed in this paper. In this framework, long-term and short-term memory networks are used to extract the underlying features in fruit fly fine-grained images. By inputting the underlying features in the multi-channel self-attention mechanism module, the global and local attention feature maps can be obtained.The weighted attention feature map can also be obtained by multiplying the weight of each channel and the attention feature map. The fine-grained image features of fruit flies were obtained by summing the weighted attention feature map. A softmax classifier was used to process the features and complete the recognition of the fruit fly fine-grained images. Two fine-grained image datasets of fruit flies were applied as experimental objects. Dataset 1 and Dataset 2 contain 11,778 images and 20,580 images from 46 different categories of fruit flies, respectively. The Kappa coefficient was used as the evaluation index to identify fruit fly images with different targets using the method proposed herein. The experimental results showed that, as the number of attention channels increased, the Kappa coefficient gradually increased, suggesting an improvement in the accuracy of fine-grained image recognition. The fine-grained image features extracted by introducing a multi-channel self-attention mechanism exhibited more distinct boundaries with a small amount of overlap, demonstrating strong feature extraction capabilities. When dealing with fine-grained images with either simple or complex backgrounds, the method proposed in this paper has good performance and generalization ability. Even if the target is small and varied in shape, it can still achieve highly accurate recognition.

Nissimagoudar P.C., Miskin S.R., Sali V.N., J A., K R.S., K D.S., M G.H., Hongal R.S., Katwe S.V., Basawaraj, I N.C.

Arvanitis G., Stagakis N., Zacharaki E.I., Moustakas K.

Lan M., Yang D., Zhou S., Ding Y.

AbstractIn order to reduce the manual workload and reduce the maintenance cost, it is particularly important to realize automatic detection of cracks. Aiming at the problems of poor real‐time performance and low precision of traditional pavement crack detection, a crack detection method based on improved YOLOv5 one‐step target detection algorithm of convolutional neural network is proposed by using the advantages of depth learning network in target detection. The images were manually marked with LabelImg annotation software, and then the network model parameters were obtained through improving the YOLOv5 network training. Finally, the cracks are verified and detected by the established model. In addition, the precision and speed of crack detection using YOLOv3, YOLOv5s, and YOLOv5s‐attention models are compared by using Precision, Recall, and F1. After comparison, it is found that the detection precision of YOLOv5s‐attention is improved by 1.0%, F1 by 0.9%, and mAP@.5 by 1.8%.

Hsieh C., Jia H., Huang W., Hsih M.

This study proposes a deep learning method for pavement defect detection, focusing on identifying potholes and cracks. A dataset comprising 10,828 images is collected, with 8662 allocated for training, 1083 for validation, and 1083 for testing. Vehicle attitude data are categorized based on three-axis acceleration and attitude change, with 6656 (64%) for training, 1664 (16%) for validation, and 2080 (20%) for testing. The Nvidia Jetson Nano serves as the vehicle-embedded system, transmitting IMU-acquired vehicle data and GoPro-captured images over a 5G network to the server. The server recognizes two damage categories, low-risk and high-risk, storing results in MongoDB. Severe damage triggers immediate alerts to maintenance personnel, while less severe issues are recorded for scheduled maintenance. The method selects YOLOv7 among various object detection models for pavement defect detection, achieving a mAP of 93.3%, a recall rate of 87.8%, a precision of 93.2%, and a processing speed of 30–40 FPS. Bi-LSTM is then chosen for vehicle vibration data processing, yielding 77% mAP, 94.9% recall rate, and 89.8% precision. Integration of the visual and vibration results, along with vehicle speed and travel distance, results in a final recall rate of 90.2% and precision of 83.7% after field testing.

Ihsan M., Amrizal M.A., Harjoko A.

This paper introduces a video dataset for semantic segmentation of road potholes. This dataset contains 619 high-resolution videos captured in January 2023, covering locations in eight villages within the Hulu Sungai Tengah regency of South Kalimantan, Indonesia. The dataset is divided into three main folders, namely train, val, and test. The train, val, and test folders contain 372 videos for training, 124 videos for validation, and 123 videos for testing, respectively. Each of these main folders has two subfolders, ``RGB'' for the video in the RGB format and ``mask'' for the ground truth segmentation. These videos are precisely two seconds long, containing 48 frames each, and all are in MP4 format. The dataset offers remarkable flexibility, accommodating various research needs, from full-video segmentation to frame extraction. It enables researchers to create ground truth annotations and change the combination of videos in the folders according to their needs. This resource is an asset for researchers, engineers, policymakers, and anyone interested in advancing algorithms for pothole detection and analysis. This dataset allows for benchmarking semantic segmentation algorithms, conducting comparative studies on pothole detection methods, and exploring innovative approaches, offering valuable contributions to the computer vision community.

Gu Y., Liu Y., Liu D., Han L.D., Jia X.

Identifying and fixing roadway potholes promptly can improve roadway safety and decrease vehicular damage, which is essential in transportation infrastructure management. Transportation agencies are constantly looking for efficient solutions for pavement intelligent management and maintenance. Traditionally, laser scanning, accelerometers, and videos are the major detecting sources employed to diagnose pavement distress. However, the responsiveness of these methods is constrained by the limited amount of equipment and labor. By contrast, popular crowdsourced tools like the navigation app Waze provide citizens with a cost-effective way to feed real-time roadway information, which could facilitate the evaluation, detection, and repair of potholes. To examine these possibilities, the study collected one-year pothole reports for five critical corridors in Nashville, Tennessee as a case study. Firstly, a Spatial Temporal Kernel Density Estimation (STKDE) was proposed to estimate the likelihood of potholes over space and time. The results highlight the spatiotemporal variation of pothole propensity. Furthermore, the hotspots recognized at a 95% confidence level revealed the significantly vulnerable areas and times, such as I-40 Westbound near log mile marker 13 from January to June 2021. Secondly, based on the optimal bandwidths generated by STKDE, spatiotemporal DBSCAN clustering was performed to identify potholes. Two pothole patterns are identified and interpreted as isolated spots and pothole zones, respectively. It was also found that crowdsourced reports come up much sooner than work requests in the Pavement Management System (PMS). Moreover, pothole reports also identified several blind spots ignored by the maintenance team. This study contributes to facilitating pavement management and maintenance with emerging crowdsourced data.

Kim G., Kim S.

We propose a novel approach to detecting road defects by leveraging smartphones. This approach presents an automatic data collection mechanism and a deep learning model for road defect detection on smartphones. The automatic data collection mechanism provides a practical and reliable way to collect and label data for road defect detection research, significantly facilitating the execution of investigations in this research field. By leveraging the automatically collected data, we designed a CNN-based model to classify speed bumps, manholes, and potholes, which outperforms conventional models in both accuracy and processing speed. The proposed system represents a highly practical and scalable technology that can be implemented using commercial smartphones, thereby presenting substantial promise for real-world applications.

Buana C., Widyastuti H., Prastyanto C.A., Rahardhita Widyatra S.

Srivani B., Kamala C., Deepti S.R., Aakash G.

Vinodhini K.A., Sidhaarth K.R.

Road surfaces are highly affected by climatic changes which caused potholes and cracks. Maintenance of the road is a need-of-the-hour process for preventing the physical damage caused for vehicles. The important process in road maintenance is the detection of potholes and cracks. Automatic detection of potholes in bituminous roads is a tedious task. This paper proposed the detection of potholes using transfer learning and convolution neural networks. The results are promising, and The suggested method can provide valuable information that can be used for various ITS services. One such service is alerting drivers about potholes, allowing them to be more cautious while driving. Additionally, this information can be utilized to assess the initial maintenance needs of a road management system and promptly address any repairs or maintenance required. The achieved results through the proposed method are compared with the state-of-the-art detection algorithms like Transfer Learning + Recurrent neural network, Transfer Learning + Generated adversarial network. In that, the result obtained through the proposed method (Transfer Learning + Convolutional neural networks achieves 96 % of accuracy.

Bibi A., Ali K., Raza A., Kausar S.

Chen X., Ma N., Wang M.

Ozoglu F., Gökgöz T.

In the context of road transportation, detecting road surface irregularities, particularly potholes, is of paramount importance due to their implications for driving comfort, transportation costs, and potential accidents. This study presents the development of a system for pothole detection using vibration sensors and the Global Positioning System (GPS) integrated within smartphones, without the need for additional onboard devices in vehicles incurring extra costs. In the realm of vibration-based road anomaly detection, a novel approach employing convolutional neural networks (CNNs) is introduced, breaking new ground in this field. An iOS-based application was designed for the acquisition and transmission of road vibration data using the built-in three-axis accelerometer and gyroscope of smartphones. Analog road data were transformed into pixel-based visuals, and various CNN models with different layer configurations were developed. The CNN models achieved a commendable accuracy rate of 93.24% and a low loss value of 0.2948 during validation, demonstrating their effectiveness in pothole detection. To evaluate the performance further, a two-stage validation process was conducted. In the first stage, the potholes along predefined routes were classified based on the labeled results generated by the CNN model. In the second stage, observations and detections during the field study were used to identify road potholes along the same routes. Supported by the field study results, the proposed method successfully detected road potholes with an accuracy ranging from 80% to 87%, depending on the specific route.

Diao Z., Huang X., Liu H., Liu Z.

Road damage detection is very important for road safety and timely repair. The previous detection methods mainly rely on humans or large machines, which are costly and inefficient. Existing algorithms are computationally expensive and difficult to arrange in edge detection devices. To solve this problem, we propose a lightweight and efficient road damage detection algorithm LE-YOLOv5 based on YOLOv5. We propose a global shuffle attention module to improve the shortcomings of the SE attention module in MobileNetV3, which in turn builds a better backbone feature extraction network. It greatly reduces the parameters and GFLOPS of the model while increasing the computational speed. To construct a simple and efficient neck network, a lightweight hybrid convolution is introduced into the neck network to replace the standard convolution. Meanwhile, we introduce the lightweight coordinate attention module into the cross-stage partial network module that was designed using the one-time aggregation method. Specifically, we propose a parameter-free attentional feature fusion (PAFF) module, which significantly enhances the model’s ability to capture contextual information at a long distance by guiding and enhancing correlation learning between the channel direction and spatial direction without introducing additional parameters. The K-means clustering algorithm is used to make the anchor boxes more suitable for the dataset. Finally, we use a label smoothing algorithm to improve the generalization ability of the model. The experimental results show that the LE-YOLOv5 proposed in this document can stably and effectively detect road damage. Compared to YOLOv5s, LE-YOLOv5 reduces the parameters by 52.6% and reduces the GFLOPS by 57.0%. However, notably, the mean average precision (mAP) of our model improves by 5.3%. This means that LE-YOLOv5 is much more lightweight while still providing excellent performance. We set up visualization experiments for multialgorithm comparative detection in a variety of complex road environments. The experimental results show that LE-YOLOv5 exhibits excellent robustness and reliability in complex road environments.

Chen Y., Wang Z., Zhang S.