An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial Vehicles
Recent advancements in the field of machine learning (ML) provide opportunity to conduct research on autonomous devices for a variety of applications. Intelligent decision-making is a critical task for self-driving systems. An attempt is made in this study to use a deep learning (DL) approach for th...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-11-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/10/22/2764 |
| _version_ | 1797510571518590976 |
|---|---|
| author | Syed-Ali Hassan Tariq Rahim Soo-Young Shin |
| author_facet | Syed-Ali Hassan Tariq Rahim Soo-Young Shin |
| author_sort | Syed-Ali Hassan |
| collection | DOAJ |
| description | Recent advancements in the field of machine learning (ML) provide opportunity to conduct research on autonomous devices for a variety of applications. Intelligent decision-making is a critical task for self-driving systems. An attempt is made in this study to use a deep learning (DL) approach for the early detection of road cracks, potholes, and the yellow lane. The accuracy is not sufficient after training with the default model. To enhance accuracy, a convolutional neural network (CNN) model with 13 convolutional layers, a softmax layer as an output layer, and two fully connected layers (FCN) are constructed. In order to achieve the deeper propagation and to prevent saturation in the training phase, mish activation is employed in the first 12 layers with a rectified linear unit (ReLU) activation function. The upgraded CNN model performs better than the default CNN model in terms of accuracy. For the varied situation, a revised and enriched dataset for road cracks, potholes, and the yellow lane is created. The yellow lane is detected and tracked in order to move the unmanned aerial vehicle (UAV) autonomously by following yellow lane. After identifying a yellow lane, the UAV performs autonomous navigation while concurrently detecting road cracks and potholes using the robot operating system within the UAV. The performance model is benchmarked using performance measures, such as accuracy, sensitivity, <i>F</i>1-<i>score</i>, <i>F</i>2-<i>score</i>, and dice-coefficient, which demonstrate that the suggested technique produces better outcomes. |
| first_indexed | 2024-03-10T05:33:14Z |
| format | Article |
| id | doaj.art-39613531cd9b4d27a2351df1484397b9 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-10T05:33:14Z |
| publishDate | 2021-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-39613531cd9b4d27a2351df1484397b92023-11-22T23:06:47ZengMDPI AGElectronics2079-92922021-11-011022276410.3390/electronics10222764An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial VehiclesSyed-Ali Hassan0Tariq Rahim1Soo-Young Shin2Department of IT Convergence Engineering, Kumoh National Institute of Technology, Gumi 39177, Gyeongbuk, KoreaDepartment of IT Convergence Engineering, Kumoh National Institute of Technology, Gumi 39177, Gyeongbuk, KoreaDepartment of IT Convergence Engineering, Kumoh National Institute of Technology, Gumi 39177, Gyeongbuk, KoreaRecent advancements in the field of machine learning (ML) provide opportunity to conduct research on autonomous devices for a variety of applications. Intelligent decision-making is a critical task for self-driving systems. An attempt is made in this study to use a deep learning (DL) approach for the early detection of road cracks, potholes, and the yellow lane. The accuracy is not sufficient after training with the default model. To enhance accuracy, a convolutional neural network (CNN) model with 13 convolutional layers, a softmax layer as an output layer, and two fully connected layers (FCN) are constructed. In order to achieve the deeper propagation and to prevent saturation in the training phase, mish activation is employed in the first 12 layers with a rectified linear unit (ReLU) activation function. The upgraded CNN model performs better than the default CNN model in terms of accuracy. For the varied situation, a revised and enriched dataset for road cracks, potholes, and the yellow lane is created. The yellow lane is detected and tracked in order to move the unmanned aerial vehicle (UAV) autonomously by following yellow lane. After identifying a yellow lane, the UAV performs autonomous navigation while concurrently detecting road cracks and potholes using the robot operating system within the UAV. The performance model is benchmarked using performance measures, such as accuracy, sensitivity, <i>F</i>1-<i>score</i>, <i>F</i>2-<i>score</i>, and dice-coefficient, which demonstrate that the suggested technique produces better outcomes.https://www.mdpi.com/2079-9292/10/22/2764autonomous navigationautonomous road inspectioncomputer visiondronerobotsneural network |
| spellingShingle | Syed-Ali Hassan Tariq Rahim Soo-Young Shin An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial Vehicles Electronics autonomous navigation autonomous road inspection computer vision drone robots neural network |
| title | An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial Vehicles |
| title_full | An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial Vehicles |
| title_fullStr | An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial Vehicles |
| title_full_unstemmed | An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial Vehicles |
| title_short | An Improved Deep Convolutional Neural Network-Based Autonomous Road Inspection Scheme Using Unmanned Aerial Vehicles |
| title_sort | improved deep convolutional neural network based autonomous road inspection scheme using unmanned aerial vehicles |
| topic | autonomous navigation autonomous road inspection computer vision drone robots neural network |
| url | https://www.mdpi.com/2079-9292/10/22/2764 |
| work_keys_str_mv | AT syedalihassan animproveddeepconvolutionalneuralnetworkbasedautonomousroadinspectionschemeusingunmannedaerialvehicles AT tariqrahim animproveddeepconvolutionalneuralnetworkbasedautonomousroadinspectionschemeusingunmannedaerialvehicles AT sooyoungshin animproveddeepconvolutionalneuralnetworkbasedautonomousroadinspectionschemeusingunmannedaerialvehicles AT syedalihassan improveddeepconvolutionalneuralnetworkbasedautonomousroadinspectionschemeusingunmannedaerialvehicles AT tariqrahim improveddeepconvolutionalneuralnetworkbasedautonomousroadinspectionschemeusingunmannedaerialvehicles AT sooyoungshin improveddeepconvolutionalneuralnetworkbasedautonomousroadinspectionschemeusingunmannedaerialvehicles |