Underwater Image Acquisition Using Raspberry Pi and Image Enhancement with Azure AI services

Introduction

Underwater image acquisition is challenging due to low visibility, poor lighting, and color distortion caused by water absorption and scattering. By integrating Raspberry Pi with a camera module for image capture and leveraging Azure Cloud for post-processing, we can enhance underwater images to improve visibility and extract meaningful insights.

This guide covers the implementation steps for underwater image acquisition using Raspberry Pi and post-processing using Azure AI services to enhance image clarity.

Start your career on Azure without leaving your job! Get Certified in less than a Month

Experienced Authorized Instructor led Training
Live Hands-on Labs

Subscribe now

1. Hardware Setup for Image Acquisition

1.1 Required Components

Raspberry Pi 4/5 – Main computing unit
Raspberry Pi Camera Module (High-Resolution, Night Vision Compatible)
Underwater Housing – Waterproof enclosure for the camera
LED Lights – To improve illumination in murky water
Battery Pack – Powering the system for underwater deployment
ROV (Optional) – If using a remotely operated vehicle for underwater navigation

2. Setting Up Image Capture on Raspberry Pi

2.1 Install Camera Dependencies

First, enable the camera module and install required software:

bash

sudo raspi-config  # Enable Camera Module

sudo apt update &amp;&amp; sudo apt upgrade -y

sudo apt install python3-opencv libopencv-dev

1

2

3

4

5

sudo raspi-config # Enable Camera Module

sudo apt update && sudo apt upgrade -y

sudo apt install python3-opencv libopencv-dev

2.2 Python Script for Image Capture

Create a Python script (capture.py) to capture and store underwater images.

python

<strong>import cv2</strong>

<strong>from picamera2 import Picamera2</strong>

<strong> </strong>

<strong># Initialize camera</strong>

<strong>picam2 = Picamera2()</strong>

<strong>picam2.preview_configuration.main.size = (1920, 1080)</strong>

<strong>picam2.preview_configuration.main.format = "RGB888"</strong>

<strong>picam2.preview_configuration.controls.FrameRate = 30</strong>

<strong>picam2.configure("preview")</strong>

<strong>picam2.start()</strong>

<strong> </strong>

<strong># Capture image</strong>

<strong>image_path = "/home/pi/underwater_image.jpg"</strong>

<strong>picam2.capture_file(image_path)</strong>

<strong>print(f"Image saved at {image_path}")</strong>

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

import cv2

from picamera2 import Picamera2

# Initialize camera

picam2 = Picamera2()

picam2.preview_configuration.main.size = (1920, 1080)

picam2.preview_configuration.main.format = "RGB888"

picam2.preview_configuration.controls.FrameRate = 30

picam2.configure("preview")

picam2.start()

# Capture image

image_path = "/home/pi/underwater_image.jpg"

picam2.capture_file(image_path)

print(f"Image saved at {image_path}")

picam2.close()

This script initializes the camera and captures an image.
The image is stored locally before being sent to Azure Cloud.

3. Uploading Images to Azure Blob Storage

3.1 Install Azure SDK on Raspberry Pi

<strong>pip3 install azure-storage-blob</strong>

1	<strong>pip3 install azure-storage-blob</strong>

3.2 Upload Image to Azure Blob Storage

Create an upload_to_azure.py script to send images to Azure Blob Storage:

<strong>from azure.storage.blob import BlobServiceClient</strong>

<strong>import os</strong>

1

2

3

from azure.storage.blob import BlobServiceClient

import os

# Azure Storage Account Details

<strong>CONNECTION_STRING = "your_azure_storage_connection_string"</strong>

<strong>CONTAINER_NAME = "underwater-images"</strong>

<strong> </strong>

<strong>def upload_image(file_path):</strong>

<strong>    blob_service_client = BlobServiceClient.from_connection_string(CONNECTION_STRING)</strong>

<strong>    blob_client = blob_service_client.get_blob_client(container=CONTAINER_NAME, blob=os.path.basename(file_path))</strong>

<strong>    with open(file_path, "rb") as data:</strong>

<strong>        blob_client.upload_blob(data, overwrite=True)</strong>

<strong>    print(f"Image {file_path} uploaded to Azure Blob Storage.")</strong>

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

CONNECTION_STRING = "your_azure_storage_connection_string"

CONTAINER_NAME = "underwater-images"

def upload_image(file_path):

blob_service_client = BlobServiceClient.from_connection_string(CONNECTION_STRING)

blob_client = blob_service_client.get_blob_client(container=CONTAINER_NAME, blob=os.path.basename(file_path))

with open(file_path, "rb") as data:

blob_client.upload_blob(data, overwrite=True)

print(f"Image {file_path} uploaded to Azure Blob Storage.")

# Upload captured image

<strong>upload_image("/home/pi/underwater_image.jpg")</strong>

1	<strong>upload_image("/home/pi/underwater_image.jpg")</strong>

This script connects to Azure Blob Storage and uploads the captured image.

4. Post-Processing for Image Enhancement in Azure

Azure provides AI-powered image processing tools to enhance underwater images:

4.1 Using Azure Computer Vision for Image Enhancement

Azure Computer Vision can improve underwater image quality by adjusting contrast, noise reduction, and color correction.

4.1.1 Setting Up Azure Cognitive Services

Go to Azure Portal → Create Cognitive Services account.
Enable Computer Vision API.
Get the API Key and Endpoint.

4.1.2 Apply Image Enhancements Using Azure AI

<strong>import requests</strong>

<strong>import json</strong>

1

2

3

import requests

import json

# Azure Computer Vision API

<strong>subscription_key = "your_subscription_key"</strong>

<strong>endpoint = "https://your-region.cognitiveservices.azure.com/vision/v3.2/analyze"</strong>

1

2

3

subscription_key = "your_subscription_key"

endpoint = "https://your-region.cognitiveservices.azure.com/vision/v3.2/analyze"

# Image URL from Azure Blob Storage

<strong>image_url = "https://yourstorageaccount.blob.core.windows.net/underwater-images/underwater_image.jpg"</strong>

<strong> </strong>

<strong>headers = {'Ocp-Apim-Subscription-Key': subscription_key, 'Content-Type': 'application/json'}</strong>

<strong>params = {'visualFeatures': 'Color,ImageType'}</strong>

<strong>data = {'url': image_url}</strong>

<strong> </strong>

<strong>response = requests.post(endpoint, headers=headers, params=params, json=data)</strong>

<strong>result = response.json()</strong>

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

image_url = "https://yourstorageaccount.blob.core.windows.net/underwater-images/underwater_image.jpg"

headers = {'Ocp-Apim-Subscription-Key': subscription_key, 'Content-Type': 'application/json'}

params = {'visualFeatures': 'Color,ImageType'}

data = {'url': image_url}

response = requests.post(endpoint, headers=headers, params=params, json=data)

result = response.json()

# Print results

<strong>print(json.dumps(result, indent=4))</strong>

1	<strong>print(json.dumps(result, indent=4))</strong>

This script analyzes the image and extracts color enhancement and noise reduction features.

5. Advanced Image Processing with Azure Machine Learning

For deeper image enhancement and noise reduction, we can use Azure Machine Learning (AML) with deep learning models:

5.1 Image Enhancement Using Deep Learning

Train a Deep Learning Model (CNN or GAN) in Azure ML Studio to improve underwater image clarity.
Deploy the trained model as an Azure ML Endpoint.
Send the captured images to the ML model for enhancement.

5.2 Python Code for Sending Image to Azure ML Model

<strong>import requests</strong>

1	<strong>import requests</strong>

# Azure ML Endpoint URL

<strong>ML_ENDPOINT = "https://your-ml-endpoint.azurewebsites.net/score"</strong>

<strong>API_KEY = "your_ml_api_key"</strong>

1

2

3

ML_ENDPOINT = "https://your-ml-endpoint.azurewebsites.net/score"

API_KEY = "your_ml_api_key"

# Send image to ML model for enhancement

<strong>headers = {'Authorization': f'Bearer {API_KEY}', 'Content-Type': 'application/json'}</strong>

<strong>payload = {'image_url': image_url}</strong>

<strong>response = requests.post(ML_ENDPOINT, json=payload, headers=headers)</strong>

<strong>enhanced_image_url = response.json().get("enhanced_image_url")</strong>

<strong>print(f"Enhanced image available at: {enhanced_image_url}")</strong>

1

2

3

4

5

6

7

8

9

headers = {'Authorization': f'Bearer {API_KEY}', 'Content-Type': 'application/json'}

payload = {'image_url': image_url}

response = requests.post(ML_ENDPOINT, json=payload, headers=headers)

enhanced_image_url = response.json().get("enhanced_image_url")

print(f"Enhanced image available at: {enhanced_image_url}")

This sends an image to Azure ML, processes it using a deep learning model, and returns an enhanced version.

6. Visualization & Monitoring in Power BI

Power BI can be used to visualize and analyze the enhanced images.
Azure Blob Storage can be connected to Power BI to track image quality over time.

7. Final Deployment Steps

Test Camera Module in different underwater conditions.
Automate Uploading Process using cron jobs on Raspberry Pi.
Optimize Image Processing using AI-based color correction for better results.

8. Future Enhancements

AI-Based Object Detection – Identifying marine life from captured images.
Real-time Image Streaming – Live feed from underwater camera using Azure Edge AI.
Integration with IoT Sensors – Combining image data with pH, salinity, and temperature readings for better insights.

Conclusion

By leveraging Raspberry Pi for underwater image acquisition and Azure AI for post-processing, we can significantly improve underwater images for research, marine studies, and environmental monitoring.

This approach enables high-quality underwater exploration with minimal hardware cost while utilizing cloud-based AI enhancements for superior image clarity.

About CloudThat

CloudThat is an award-winning company and the first in India to offer cloud training and consulting services worldwide. As a Microsoft Solutions Partner, AWS Advanced Tier Training Partner, and Google Cloud Platform Partner, CloudThat has empowered over 850,000 professionals through 600+ cloud certifications winning global recognition for its training excellence including 20 MCT Trainers in Microsoft’s Global Top 100 and an impressive 12 awards in the last 8 years. CloudThat specializes in Cloud Migration, Data Platforms, DevOps, IoT, and cutting-edge technologies like Gen AI & AI/ML. It has delivered over 500 consulting projects for 250+ organizations in 30+ countries as it continues to empower professionals and enterprises to thrive in the digital-first world.