TriView-YOLO: Early Multi-View Fusion for Ground Penetrating Radar Cavity Detection in Soft, High-Water-Content Soils
2026-08-10 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed a new method called TriView-YOLO to find underground cavities in soft, waterlogged soil using special radar images. They combined three types of radar views into one input to improve detection, focusing on road cavities in challenging environments like Bangkok's wet clay soils. Their system was trained on real field data and tested specifically on these difficult conditions, achieving moderate accuracy. They found that using multiple views helped, while adding synthetic data or larger models did not improve results.
Ground Penetrating Radar (GPR)subsurface cavity detectionYOLOv12multiview inputB-scanC-scanmarine clay soilmean Average Precision (mAP)deep learningsignal attenuation
Authors
Suphawut Thawinutchokaudom, Sompote Youwai, Warat Kongkitkul, Mitsumasa Yamashina, Jose M. D. S. Rodrigues Neto, Jun Shinohara
Abstract
Automated detection of subsurface cavities from Ground Penetrating Radar (GPR) is most difficult in soft, high-water-content ground, where conductive, water-saturated soil attenuates the signal and degrades cavity reflections, yet this is also the condition under which cavities most readily form. This paper proposes TriView-YOLO, a multi-view YOLOv12 detector for road cavity screening in such ground. Three co-registered views (longitudinal B-scan, horizontal C-scan, and cross-section B-scan) form a 9-channel input fused by a TripleInputConv layer that replaces the YOLOv12 stem; the rest of the network is unchanged, and bounding boxes are required on the longitudinal view only. Training used 1,600 expert-verified field samples, principally metropolitan road surveys of Bangkok, Thailand, acquired with a vehicle-mounted multichannel three-dimensional GPR mobile mapping system, with surveys over the firmer subgrades of Japan added to training and validation only. The test set comes exclusively from the Bangkok surveys, over soft marine clay with 80-140% water content and a water table at 1-2 m depth, a ground condition for which no dedicated deep learning cavity-detection evaluation has been reported. On this unaugmented, field-only test set, split randomly within surveys, the proposed model attains mAP50 of 0.558 +/- 0.028 over three seeds at 23.6 GFLOPs and 3.1 ms per image. Ablations show that removing the auxiliary views lowers mAP50 and recall, whereas public and synthetic training images, DINOv3 features, larger model scale, and COCO pretraining bring no gain.