---
title: "Water Meters Dataset"
description: "2,488+ photos"
url: "https://unidata.pro/datasets/water-meters/"
date_modified: "2025-12-24T15:48:58+03:00"
language: "en-US"
---
It is a collection of water meter images with segmentation masks and OCR-labeled meter readings, designed to support detection and meter reading tasks in AI models for monitoring water usage, improving customer service, and enhancing wastewater treatment and water supply systems.

## Dataset Structure

### The Numbers Section

**Numbered list:** - **Number:** 2,488+ — **Text:** photos

### Tooltips Section

**Tooltip items:**

- **Name:** Detection
- **Name:** Computer Vision
- **Name:** Smart city
- **Name:** Communal services
- **Name:** Machine learning

### Dataset Information

**Table with data:**

| Characteristic | Data |
| --- | --- |
| Description | Images of water meters, including their segmentation masks and OCR labels for meter readings |
| Data types | Image |
| Tasks | Meter readings, Detection |
| Total number of images | 2,488 |
| Number sets | 1,244 |
| Number of files in a set | 2 (photo + mask) |
| Marking | OCR |
| Labeling | value of the water meter, location of the bounding box |

**Media Slider:**

- **Image in the slider:** ![Labeling of the image without crop](https://unidata.pro/wp-content/uploads/2024/12/water-meters-labeling-of-the-image-without-crop.webp)
- **Image in the slider:** ![Original image](https://unidata.pro/wp-content/uploads/2024/12/water-meters-original-image-rotated.webp)

**Link to the sample:** [Download sample](https://drive.google.com/drive/u/0/folders/18UMXLWPYnHTnGbjmwSutPuWjKZCHdXMW)

### Technical Specifications

**Table with data:**

| Characteristic | Data |
| --- | --- |
| Image extension | JPG |
| Extension of labeling file | CSV |

**Source and data collection methodology:** Source and collection methodology. Data was collected via crowdsourcing platforms.

### Dataset Use Cases - Slider

**Industry Cards:**

- **Industry:** Utility Management — **Title:** Automating Water Meter Reading Systems — **Text:** Water Meters Dataset provides high-quality water meter pics that help utility companies automate the process of meter reading. By training computer vision models to recognize numbers and dial positions on meters, organizations can reduce manual labor, minimize reading errors, and improve billing accuracy across various meter sizes and conditions.
- **Industry:** Smart Infrastructure Development — **Title:** Building Intelligent Monitoring Solutions — **Text:** This dataset supports the development of smart metering systems that continuously track water usage and detect anomalies in real time. Using images of water meters collected under different lighting and environmental conditions, engineers can create AI-powered solutions that improve water resource management, identify leaks, and optimize consumption reporting for both residential and commercial users.
- **Industry:** Water Utility Operations — **Title:** Enhancing Customer Service and Billing Systems — **Text:** Utility providers use the dataset to train machine learning models for automated customer billing and usage verification. The diversity of water meter pics enables the creation of reliable image recognition tools that read meters accurately, even in outdoor or worn conditions, supporting efficient wastewater services, timely reporting, and better customer satisfaction.
- **Industry:** Environmental Monitoring and Research — **Title:** Supporting Sustainable Water Management Initiatives — **Text:** Researchers can apply Water Meters Dataset to analyze water consumption trends and develop predictive models for efficient water supply management. By combining accurate meter readings with other water quality and usage data, it contributes to smarter resource planning, reduced waste, and the development of sustainable drinking water and wastewater treatment systems.

### Fact

**FAQs Heading:** FAQs

**List of Questions:**

- **Question:** What is this dataset used for? — **Answer:** Water Meters Dataset is designed for automatic meter reading (AMR), object detection, and OCR model training. It helps utility companies and AI developers build systems for meter reading automation, water consumption analysis, and billing accuracy improvement.
- **Question:** What types of annotations are provided? — **Answer:** Each water meter image includes OCR labels containing the numerical meter reading and the bounding box location.
- **Question:** Can I request a sample of the dataset before purchasing? — **Answer:** Yes. Unidata provides free dataset samples so you can assess image quality, OCR label accuracy, and segmentation mask consistency before making a purchase. This allows you to test compatibility with your meter reading recognition or AI detection models.
- **Question:** How do labeled water meter images improve OCR and detection models? — **Answer:** Labeled water meter images provide AI models with examples of meter displays and visual features, helping them accurately locate meters and extract numerical readings. This improves the reliability of automated recognition systems in real-world conditions.
- **Question:** What are the sources of data for Unidata datasets? — **Answer:** All Unidata datasets, including Water Meters Dataset, are collected via crowdsourcing platforms. Each image is verified and annotated by trained contributors, ensuring high-quality data labeling and diverse meter reading samples across various water meter types and sizes.
- **Question:** How are Unidata datasets licensed? — **Answer:** Unidata datasets follow a dual-licensing model. Free samples are available for testing and validation, while the full dataset can be purchased for commercial, research, or industrial use.
- **Question:** Do Unidata datasets follow GDPR or other data privacy regulations? — **Answer:** Yes. All datasets are fully compliant with GDPR and applicable data privacy regulations. Water Meters Dataset contains no personally identifiable information, and all data collection is performed through lawful and ethical means.
- **Question:** How are Unidata datasets stored? — **Answer:** Unidata securely stores all datasets on AWS cloud infrastructure, ensuring reliability, scalability, and data protection. Storage practices align with ISO 27001 and ISO 27701 standards, guaranteeing a privacy-focused environment for dataset management.
- **Question:** How long does it take to receive the dataset? — **Answer:** After you submit a request, our team will contact you to confirm the details and finalize documentation. Once signing and payment are complete, the dataset will be delivered within 3–10 business days.

[Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)