Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in Soil
Soil desiccation cracks in natural clayey soil pose significant risks to the stability of civil and geotechnical structures. Traditional methods for detecting these cracks are often inefficient and prone to inaccuracies. Therefore, we applied a deep learning approach of semantic segmentation based o...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-04-01
|
| Series: | Engineering Proceedings |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2673-4591/91/1/2 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1839654035934674944 |
|---|---|
| author | Hui Yean Ling See Hung Lau Siaw Yah Chong Min Lee Lee Yasuo Tanaka |
| author_facet | Hui Yean Ling See Hung Lau Siaw Yah Chong Min Lee Lee Yasuo Tanaka |
| author_sort | Hui Yean Ling |
| collection | DOAJ |
| description | Soil desiccation cracks in natural clayey soil pose significant risks to the stability of civil and geotechnical structures. Traditional methods for detecting these cracks are often inefficient and prone to inaccuracies. Therefore, we applied a deep learning approach of semantic segmentation based on DeepLabv3+ to detect desiccation cracks. To enhance computational efficiency, a pretrained lightweight network, MobileNetV2, was employed as the backbone for the DeepLabv3+ model. The model was trained and tested on a dataset of natural clayey soil crack images obtained through laboratory tests. Evaluation metrics including precision, recall, F1 score, and intersection over union (IoU) were used to assess the segmentation performance. The model took 17.13 min to train and achieved an inference speed of 0.43 s per image. DeepLabv3+ achieved better performance than the traditional segmentation method, with a precision of 95.76%, a recall of 84.12%, an F1 score of 89.56%, and an IoU of 81.10%. The model also demonstrated the capability to handle images with shading conditions and the presence of spots. DeepLabv3+ with MobileNetV2 as a backbone network was proven to be effective and efficient as a backbone in soil desiccation crack detection and segmentation. |
| format | Article |
| id | doaj-art-08ef8a4c53ae4151a2b727a6d579ea97 |
| institution | Matheson Library |
| issn | 2673-4591 |
| language | English |
| publishDate | 2025-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Engineering Proceedings |
| spelling | doaj-art-08ef8a4c53ae4151a2b727a6d579ea972025-06-25T13:47:56ZengMDPI AGEngineering Proceedings2673-45912025-04-01911210.3390/engproc2025091002Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in SoilHui Yean Ling0See Hung Lau1Siaw Yah Chong2Min Lee Lee3Yasuo Tanaka4Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Kajang 43000, Selangor, MalaysiaLee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Kajang 43000, Selangor, MalaysiaLee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Kajang 43000, Selangor, MalaysiaDepartment of Civil Engineering, University of Nottingham Malaysia, Semenyih 43500, Selangor, MalaysiaLee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Kajang 43000, Selangor, MalaysiaSoil desiccation cracks in natural clayey soil pose significant risks to the stability of civil and geotechnical structures. Traditional methods for detecting these cracks are often inefficient and prone to inaccuracies. Therefore, we applied a deep learning approach of semantic segmentation based on DeepLabv3+ to detect desiccation cracks. To enhance computational efficiency, a pretrained lightweight network, MobileNetV2, was employed as the backbone for the DeepLabv3+ model. The model was trained and tested on a dataset of natural clayey soil crack images obtained through laboratory tests. Evaluation metrics including precision, recall, F1 score, and intersection over union (IoU) were used to assess the segmentation performance. The model took 17.13 min to train and achieved an inference speed of 0.43 s per image. DeepLabv3+ achieved better performance than the traditional segmentation method, with a precision of 95.76%, a recall of 84.12%, an F1 score of 89.56%, and an IoU of 81.10%. The model also demonstrated the capability to handle images with shading conditions and the presence of spots. DeepLabv3+ with MobileNetV2 as a backbone network was proven to be effective and efficient as a backbone in soil desiccation crack detection and segmentation.https://www.mdpi.com/2673-4591/91/1/2semantic segmentationsoil desiccation crackcrack detectiondeep learningDeepLabv3+ |
| spellingShingle | Hui Yean Ling See Hung Lau Siaw Yah Chong Min Lee Lee Yasuo Tanaka Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in Soil Engineering Proceedings semantic segmentation soil desiccation crack crack detection deep learning DeepLabv3+ |
| title | Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in Soil |
| title_full | Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in Soil |
| title_fullStr | Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in Soil |
| title_full_unstemmed | Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in Soil |
| title_short | Semantic Segmentation Using Lightweight DeepLabv3+ for Desiccation Crack Detection in Soil |
| title_sort | semantic segmentation using lightweight deeplabv3 for desiccation crack detection in soil |
| topic | semantic segmentation soil desiccation crack crack detection deep learning DeepLabv3+ |
| url | https://www.mdpi.com/2673-4591/91/1/2 |
| work_keys_str_mv | AT huiyeanling semanticsegmentationusinglightweightdeeplabv3fordesiccationcrackdetectioninsoil AT seehunglau semanticsegmentationusinglightweightdeeplabv3fordesiccationcrackdetectioninsoil AT siawyahchong semanticsegmentationusinglightweightdeeplabv3fordesiccationcrackdetectioninsoil AT minleelee semanticsegmentationusinglightweightdeeplabv3fordesiccationcrackdetectioninsoil AT yasuotanaka semanticsegmentationusinglightweightdeeplabv3fordesiccationcrackdetectioninsoil |