Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -13,4 +13,78 @@ language:
|
|
| 13 |
pretty_name: U-Tube Manometer Detection Dataset 2025-2026
|
| 14 |
size_categories:
|
| 15 |
- 1K<n<10K
|
| 16 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
pretty_name: U-Tube Manometer Detection Dataset 2025-2026
|
| 14 |
size_categories:
|
| 15 |
- 1K<n<10K
|
| 16 |
+
---
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
<img src="https://github.com/ClaytonSdS/VisionGauge/blob/main/steps/utm_dataset_detection_logo.png?raw=true" width="900"/>
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
# **Dataset Description**
|
| 23 |
+
|
| 24 |
+
This dataset was created to support Article X and is available on [Roboflow](https://universe.roboflow.com/visiongauge/utm_dataset-ooorv) and [Hugging Face](https://huggingface.co/datasets/claytonsds/UTM_Dataset).
|
| 25 |
+
|
| 26 |
+
It consists of images of **U-tube manometers** constructed using a **transparent PVC water level hose (5/16" × 1 mm)** and **flexible measuring tapes of different colors**, each with a length of **150 cm (60 inches)**. The manometric fluids represented in the dataset include **water, oil, and dyed water**, and the dataset is intended for computer vision object detection tasks, specifically targeting the **detection of liquid column levels**.
|
| 27 |
+
|
| 28 |
+
The dataset includes **8889 images**, and all annotations were performed using the RoboFlow Annotate Tool.
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
# **Dataset Split**
|
| 32 |
+
|
| 33 |
+
* **Training set:** 6253 images (70%)
|
| 34 |
+
* **Validation set:** 1758 images (20%)
|
| 35 |
+
* **Test set:** 878 images (10%)
|
| 36 |
+
|
| 37 |
+
# Data Structure
|
| 38 |
+
|
| 39 |
+
The dataset follows a standard split structure:
|
| 40 |
+
```
|
| 41 |
+
dataset/
|
| 42 |
+
├── train/
|
| 43 |
+
│ ├── images/
|
| 44 |
+
│ │ ├── image5214.jpg
|
| 45 |
+
│ │ └── ...
|
| 46 |
+
│ └── labels/
|
| 47 |
+
│ ├── image5214.txt
|
| 48 |
+
│ └── ...
|
| 49 |
+
├── validation/
|
| 50 |
+
│ ├── images/
|
| 51 |
+
│ │ ├── image3710.jpg
|
| 52 |
+
│ │ └── ...
|
| 53 |
+
│ └── labels/
|
| 54 |
+
│ ├── image3710.txt
|
| 55 |
+
│ └── ...
|
| 56 |
+
└── test/
|
| 57 |
+
│ ├── images/
|
| 58 |
+
│ │ ├── image18313.jpg
|
| 59 |
+
│ │ └── ...
|
| 60 |
+
│ └── labels/
|
| 61 |
+
│ ├── image18313.txt
|
| 62 |
+
│ └── ...
|
| 63 |
+
└── data.yaml
|
| 64 |
+
```
|
| 65 |
+
|
| 66 |
+
* The dataset includes TXT annotations and a YAML configuration file compatible with YOLOv8.
|
| 67 |
+
|
| 68 |
+
---
|
| 69 |
+
|
| 70 |
+
# License
|
| 71 |
+
|
| 72 |
+
This dataset is released under the **Apache License 2.0**.
|
| 73 |
+
|
| 74 |
+
---
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
# **Citation**
|
| 78 |
+
|
| 79 |
+
```bibtex
|
| 80 |
+
@misc{utm_detection_dataset,
|
| 81 |
+
author = {Santos, Clayton Silva},
|
| 82 |
+
title = {{UTM} {Detection Dataset}},
|
| 83 |
+
year = {2026},
|
| 84 |
+
month = {feb},
|
| 85 |
+
publisher = {Roboflow},
|
| 86 |
+
version = {12},
|
| 87 |
+
doi = {10.57967/hf/7558},
|
| 88 |
+
url = {http://universe.roboflow.com/visiongauge/utm_dataset_detection/dataset/12}
|
| 89 |
+
}
|
| 90 |
+
```
|