TulikaC presents a Python and OpenCV-based web app for enhancing low-light images, with practical deployment instructions for Azure App Service. This guide covers the key steps in setting up the enhancement pipeline and deploying using Azure Developer CLI.

Deploying a Low-Light Image Enhancer (Python + OpenCV) on Azure App Service

Author: TulikaC

This guide shows how to build and deploy a lightweight, explainable low-light image enhancement web app using Python, Flask, and OpenCV. The solution leverages classic image processing techniques, wrapped in a convenient web interface, and is deployable on Azure App Service for Linux using the Azure Developer CLI.

Overview

  • App Purpose: Instantly enhance low-light photos in the browser using a fast, explainable pipeline (no heavyweight ML models required).
  • Pipeline Stages:
    • CLAHE (Contrast Limited Adaptive Histogram Equalization)
    • Gamma Correction
    • Brightness Adjustment
    • Subtle Saturation Boost

What You’ll Build

  • A Flask web app accepting image uploads (with tunable parameters).
  • On upload, the backend applies the enhancement pipeline and returns both original and improved images as base64 PNG data URIs—enabling side-by-side browser previews without persistent storage.

App Architecture

  1. Web Browser posts to /enhance with image and settings (clip_limit, gamma, brightness)
  2. Flask Backend converts upload to NumPy RGB and calls LowLightEnhancer.enhance_image()
  3. Response returns JSON: base64-encoded original and enhanced images

Prerequisites

  • Azure subscription
  • Azure Developer CLI (azd) installed (Setup instructions)
  • (Optional) Python 3.9+ for local development

Quick Deployment Steps

git clone https://github.com/Azure-Samples/appservice-ai-samples.git
cd appservice-ai-samples/lowlight-enhancer
azd init
azd up

Once complete, navigate to the provided URL, upload an image, and compare the enhanced results.

Key Implementation Details

1. Flask App Surface

  • File upload size limited to 16MB:

    app = Flask(__name__)
app.config['MAX_CONTENT_LENGTH'] = 16 * 1024 * 1024
  • Routes:
    • / (GET): Renders upload UI
    • /enhance (POST): Accepts uploads, returns JSON with images

  • Parameter handling with defaults:

    clip_limit = float(request.form.get('clip_limit', 2.0))
gamma = float(request.form.get('gamma', 1.2))
brightness = float(request.form.get('brightness', 1.1))

2. In-Memory Processing and Data URI Response

  • No temp files needed: Input is converted to RGB, enhanced, then both versions are base64-encoded for instant inline display. Key functions:

    def process_uploaded_image(file_storage, clip_limit=2.0, gamma=1.2, brightness=1.1):
    img_pil = Image.open(file_storage)
    if img_pil.mode != 'RGB':
        img_pil = img_pil.convert('RGB')
    img_array = np.array(img_pil)
    enhanced = LowLightEnhancer().enhance_image(
        img_array,
        clip_limit=clip_limit,
        gamma=gamma,
        brightness_boost=brightness
    )
    # Convert both to base64 PNGs for response
    ...

3. Enhancement Pipeline (enhancer.py)

  • CLAHE applied to L-channel in LAB color space (prevents color artifacts)
  • Gamma correction with a lookup table for perceptual brightness
  • Brightness boost is a multiplier post-CLAHE/gamma

  • Saturation: A 10% bump restores color vibrancy

    class LowLightEnhancer:
    def enhance_image(self, image, ...):
        ...
        # Apply CLAHE on LAB L channel
        ...
        # Gamma correction
        ...
        # Brightness lift
        ...
        # Merge and add subtle saturation boost
        ...
        return enhanced

4. Tuning Tips

  • Too flat/muddy? Raise clip_limit (e.g., 3.0–4.0)
  • Too dark? Increase gamma (e.g., 1.4–1.6)
  • Too harsh? Lower clip_limit or decrease brightness_boost
  • Color too muted? Raise saturation factor slightly (e.g., 1.15)

5. Next Steps and Enhancements

  • Add UI controls for more parameters (e.g., tile grid size)
  • Store images in Azure Blob Storage for sharing
  • Enable batch processing via background jobs

Resources

For further learning, explore more Azure App Service samples and best practices provided in the official documentation and GitHub repository.

This post appeared first on “Microsoft Tech Community”. Read the entire article here