Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection
To ensure the safe operation of highway traffic lines, given the imperfect feature extraction of existing road pit defect detection models and the practicability of detection equipment, this paper proposes a lightweight target detection algorithm with enhanced feature extraction based on the YOLO (Y...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-05-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/22/9/3537 |
_version_ | 1797502712975196160 |
---|---|
author | Fu-Jun Du Shuang-Jian Jiao |
author_facet | Fu-Jun Du Shuang-Jian Jiao |
author_sort | Fu-Jun Du |
collection | DOAJ |
description | To ensure the safe operation of highway traffic lines, given the imperfect feature extraction of existing road pit defect detection models and the practicability of detection equipment, this paper proposes a lightweight target detection algorithm with enhanced feature extraction based on the YOLO (You Only Look Once) algorithm. The BIFPN (Bidirectional Feature Pyramid Network) network structure is used for multi-scale feature fusion to enhance the feature extraction ability, and Varifocal Loss is used to optimize the sample imbalance problem, which improves the accuracy of road defect target detection. In the evaluation test of the model in the constructed PCD1 (Pavement Check Dataset) dataset, the mAP@.5 (mean Average Precision when IoU = 0.5) of the BV-YOLOv5S (BiFPN Varifocal Loss-YOLOv5S) model increased by 4.1%, 3%, and 0.9%, respectively, compared with the YOLOv3-tiny, YOLOv5S, and B-YOLOv5S (BiFPN-YOLOv5S; BV-YOLOv5S does not use the Improved Focal Loss function) models. Through the analysis and comparison of experimental results, it is proved that the proposed BV-YOLOv5S network model performs better and is more reliable in the detection of pavement defects and can meet the needs of road safety detection projects with high real-time and flexibility requirements. |
first_indexed | 2024-03-10T03:40:01Z |
format | Article |
id | doaj.art-5fb62647b2e4415ab305453d73353bf1 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-10T03:40:01Z |
publishDate | 2022-05-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-5fb62647b2e4415ab305453d73353bf12023-11-23T09:19:55ZengMDPI AGSensors1424-82202022-05-01229353710.3390/s22093537Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect DetectionFu-Jun Du0Shuang-Jian Jiao1Department of Civil Engineering, College of Engineering, Ocean University of China, Qingdao 266100, ChinaDepartment of Civil Engineering, College of Engineering, Ocean University of China, Qingdao 266100, ChinaTo ensure the safe operation of highway traffic lines, given the imperfect feature extraction of existing road pit defect detection models and the practicability of detection equipment, this paper proposes a lightweight target detection algorithm with enhanced feature extraction based on the YOLO (You Only Look Once) algorithm. The BIFPN (Bidirectional Feature Pyramid Network) network structure is used for multi-scale feature fusion to enhance the feature extraction ability, and Varifocal Loss is used to optimize the sample imbalance problem, which improves the accuracy of road defect target detection. In the evaluation test of the model in the constructed PCD1 (Pavement Check Dataset) dataset, the mAP@.5 (mean Average Precision when IoU = 0.5) of the BV-YOLOv5S (BiFPN Varifocal Loss-YOLOv5S) model increased by 4.1%, 3%, and 0.9%, respectively, compared with the YOLOv3-tiny, YOLOv5S, and B-YOLOv5S (BiFPN-YOLOv5S; BV-YOLOv5S does not use the Improved Focal Loss function) models. Through the analysis and comparison of experimental results, it is proved that the proposed BV-YOLOv5S network model performs better and is more reliable in the detection of pavement defects and can meet the needs of road safety detection projects with high real-time and flexibility requirements.https://www.mdpi.com/1424-8220/22/9/3537pavement defectsdeep learningconvolutional neural networkYOLOv5Sautomated inspectionembedded equipment |
spellingShingle | Fu-Jun Du Shuang-Jian Jiao Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection Sensors pavement defects deep learning convolutional neural network YOLOv5S automated inspection embedded equipment |
title | Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection |
title_full | Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection |
title_fullStr | Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection |
title_full_unstemmed | Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection |
title_short | Improvement of Lightweight Convolutional Neural Network Model Based on YOLO Algorithm and Its Research in Pavement Defect Detection |
title_sort | improvement of lightweight convolutional neural network model based on yolo algorithm and its research in pavement defect detection |
topic | pavement defects deep learning convolutional neural network YOLOv5S automated inspection embedded equipment |
url | https://www.mdpi.com/1424-8220/22/9/3537 |
work_keys_str_mv | AT fujundu improvementoflightweightconvolutionalneuralnetworkmodelbasedonyoloalgorithmanditsresearchinpavementdefectdetection AT shuangjianjiao improvementoflightweightconvolutionalneuralnetworkmodelbasedonyoloalgorithmanditsresearchinpavementdefectdetection |