claytonsds commited on
Commit
6643dda
·
verified ·
1 Parent(s): e1b5afe

update: README.md

Browse files
Files changed (1) hide show
  1. README.md +125 -10
README.md CHANGED
@@ -1,17 +1,132 @@
1
  ---
2
  license: apache-2.0
 
3
  task_categories:
4
- - image-classification
5
- - object-detection
 
6
  language:
7
- - en
8
- - pt
 
9
  tags:
10
- - image-regression
11
- - u-tube-manometer
12
- - gauge-reader
13
- - water-column
 
 
14
  pretty_name: U-Tube Manometer Dataset 2025-2026
 
15
  size_categories:
16
- - 10K<n<100K
17
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: apache-2.0
3
+
4
  task_categories:
5
+ - image-classification
6
+ - image-regression
7
+
8
  language:
9
+ - en
10
+ - pt
11
+
12
  tags:
13
+ - image-regression
14
+ - image-classification
15
+ - u-tube-manometer
16
+ - gauge-reader
17
+ - water-column
18
+
19
  pretty_name: U-Tube Manometer Dataset 2025-2026
20
+
21
  size_categories:
22
+ - 10K<n<100K
23
+ ---
24
+
25
+ # U-Tube Manometer Dataset (2025–2026)
26
+
27
+ This dataset contains images of **U-tube manometers** and is designed for **computer vision regression and classification tasks**, focusing on reading and interpreting liquid column levels.
28
+
29
+ ---
30
+
31
+ ## Dataset Summary
32
+
33
+ The **U-Tube Manometer Dataset** consists of 20,000 images of U-tube manometers with visible water columns. The dataset is suitable for:
34
+
35
+ * **Image Regression**: estimating the numerical value of the water column height
36
+ * **Image Classification**: classifying images based on discrete column values or ranges
37
+
38
+ Images are organized in a folder structure where the **label is encoded in the directory name**, representing the measured water column value.
39
+
40
+ ---
41
+
42
+ ## Supported Tasks
43
+
44
+ * Image Regression
45
+ * Image Classification
46
+
47
+ ---
48
+
49
+ ## Data Structure
50
+
51
+ The dataset follows a standard split structure:
52
+
53
+ ```
54
+ dataset/
55
+ ├── train/
56
+ │ ├── 10.2/
57
+ │ │ ├── image23.png
58
+ │ │ └── ...
59
+ │ ├── 10.4/
60
+ │ └── ...
61
+ ├── validation/
62
+ │ ├── 9.8/
63
+ │ └── ...
64
+ └── test/
65
+ ├── 10.0/
66
+ └── ...
67
+ ```
68
+
69
+ * Each subfolder name corresponds to the **label value** (e.g., water column height).
70
+ * Images inside the folder share the same label.
71
+
72
+ ---
73
+
74
+ ## Labels
75
+
76
+ * **Type**: Numerical (float)
77
+ * **Description**: Water column height (e.g., in centimeters or equivalent unit)
78
+ * **Source**: Encoded directly in the directory name
79
+
80
+ For classification tasks, labels can be treated as discrete classes.
81
+
82
+ ---
83
+
84
+ ## Dataset Size
85
+
86
+ * **Total images**: 20,000
87
+
88
+ ---
89
+
90
+ ## Image Characteristics
91
+
92
+ * RGB images
93
+ * Varying lighting conditions
94
+ * Varying working fluids (water, oil, water–ink mixtures)
95
+ * Single U-tube manometer per image
96
+ * Visible liquid columns and scale markings
97
+
98
+ ---
99
+
100
+ ## Intended Use
101
+
102
+ This dataset is intended for:
103
+
104
+ * Training and evaluating regression models for visual measurement
105
+ * Benchmarking image-based gauge reading systems
106
+ * Research in applied computer vision for instrumentation
107
+
108
+ ---
109
+
110
+ ## Limitations
111
+
112
+ * Performance may vary under extreme lighting or occlusion conditions
113
+
114
+ ---
115
+
116
+ ## License
117
+
118
+ This dataset is released under the **Apache License 2.0**.
119
+
120
+ ---
121
+
122
+ ## Citation
123
+
124
+ If you use this dataset in your research, please cite:
125
+
126
+ ```bibtex
127
+ @dataset{u_tube_manometer_2025,
128
+ title={U-Tube Manometer Dataset},
129
+ year={2025},
130
+ license={Apache-2.0}
131
+ }
132
+ ```