CUBIT: A High-resolution Infrastructure Defect Dataset Evaluated with Autonomous Detection Framework
2.School of Electronics and Information Engineering,Tongji University
- The visualization of defect detection framework based on CUBIT dataset is illustrated below *, which encompasses the entire process: data collection by autonomous unmanned system; the baseline network integrated with our GIPFPP module; the output of defect detection results.
| Dataset | Num. of Images | Resolution | Data Collection Platform | Category | Scenario | Material | Experiments |
|---|---|---|---|---|---|---|---|
| RDD-2018 | 9053 | 600x600 | Smartphones | Crack, Corrosion | Pavement | Asphalt | SSD |
| RDD-2019 | 13135 | 600x600 | Smartphones | Crack, Corrosion | Pavement | Asphalt | SSD |
| RDD-2020 | 26336 | 600x600, 720x720 | Smartphones | Crack, Pothole | Pavement | Asphalt | SSD |
| RDD-2022 | 47420 | 512x512, 600x600, 720x720, 3650x2044 | Smartphones, Hand-held cameras, UAV cameras, Google street view | Crack, Pothole | Pavement | Asphalt | - |
| PID | 7237 | 640x640 | Crawled from Internet | Crack | Pavement | Asphalt | YOLOv2, Fast R-CNN |
| Murad | 2620 | up to 838x809 | Smartphones | Crack | Pavement | Asphalt | Faster R-CNN |
| CODEBRIM | 1590 | up to 6000x4000 | Hand-held cameras, UAV Cameras | Crack, Corrosion | Bridge | Concrete | MetaQNN, ENAS |
| CUBIT | 5527 | 4624x3472 and 8000x6000 | Cameras in Unmanned Systems | Crack, Spallinig, Moisture | Building (65%), Pavement (29%), Bridge (6%) | Concrete, Asphalt, Stone | Faster R-CNN, PP-YOLO, PP-YOLOv2, YOLOX, YOLOv5, YOLOv7, YOLOv6, YOLOv6+GIPFPP(ours), Real-site experiment |
The evaluation results of SOTA real-time detection methods and YOLOv6-n with our GIPFPP module are benchmarked in the table below. After switching from the original module to GIPFPP module, the AP of YOLOv6-n is improved by 3%, while its number of parameters is reduced by 10%. The enhancements made to the model will facilitate the real-time defect detection using unmanned systems.
| Model | #Params.(M) | FLOPs(G) | Size | mAP${50}^{test}$ / mAP${50:95}^{test}$ | Latency(ms) |
|---|---|---|---|---|---|
| Faster R-CNN(Res50) | 42.62 | 477.24 | 1024 | 71.5% / 43.3% | 76.9 |
| PP-YOLO | 48.99 | 136.43 | 1024 | 76.4% / 45.1% | 14.5 |
| PP-YOLOv2 | 56.91 | 146.50 | 1024 | 77.3% / 47.1% | 13.8 |
| YOLOv5-n | 1.76 | 4.10 | 1024 | 73.4% / 39.9% | 1.8 |
| YOLOv5-s | 7.18 | 15.80 | 1024 | 78.5% / 47.2% | 3.3 |
| YOLOv7-t | 6.01 | 13.01 | 1024 | 71.1% / 39.7% | 1.9 |
| YOLOX-n | 2.24 | 17.75 | 1024 | 73.0% / 39.5% | 4.4 |
| YOLOX-t | 5.03 | 39.00 | 1024 | 75.3% / 49.2% | 5.8 |
| YOLOX-s | 8.94 | 68.51 | 1024 | 77.9% / 49.4% | 7.6 |
| YOLOv6-n(baseline) | 4.63 | 29.03 | 1024 | 76.3% / 47.9% | 2.2 |
| YOLOv6-s | 18.50 | 115.64 | 1024 | 79.0% / 48.2% | 5.3 |
| YOLOv6-n+GIFPFF(ours) | 4.14 (-0.49) | 28.02 (-1.01) | 1024 | 77.5% (+1.2) / 50.3% (+3.1) | 2.2 |
We enlarge the prediction results in the bottom right corner of framework images above. CUBIT dataset covers three infrastructure types: Building facade, Pavement, and Bridge, and aims for three types of defect: Crack, Spalling, and Moisture. Rectangles indicate the output prediction box Red for Crack, Pink for Spalling, and Orange for Moisture with inferred defect type and confidence score from YOLOv6-l trained on the training set of our proposed dataset.
Qualitative visualization of UAV-based real-world experiment is shown below On the left, our multi-UAVs inspection schematics is illustrated. On the right, the detection results of four direction façades of the building are displayed.
