Blur Detector

😫 The Pain Point

After an event shoot, you have 500 photos but 100 are blurry. Reviewing each one to find the bad shots takes hours.

🚀 Agentic Solution

An Automatic Blur Detector that scores each image and separates the sharp from the blurry.

Key Features:

• Laplacian Variance: Mathematical measurement of image sharpness.
• Adjustable Threshold: Customize what counts as “too blurry”.
• Batch Sorting: Automatically move blurry images to a separate folder.

⚔️ Phase 1: Commander (Quick Fix)

For quick sorting.

Prompt:

“I have a folder event_photos with mixed quality images. Write a Python script using OpenCV to:

1. Blur Detection: Calculate Laplacian variance for each image.
2. Threshold: Images with variance < 100 are considered blurry (adjustable).
3. Sort: Move blurry images to blurry/ folder, keep sharp in sharp/.
4. Report: Print each image with its blur score.

Handle read errors gracefully. Print summary at end.”

Result: Only sharp photos remain for editing.

🏗️ Phase 2: Architect (Permanent Tool)

For Photographers.

Engineering Prompt:

**Role:** Python Tool Developer
**Task:** Create a "Photo Quality Sorter".

**Requirements:**
1.  **GUI:**
    *   Select folder.
    *   Threshold slider with preview.
    *   Show blur score distribution chart.
    *   Preview borderline images.
    *   Move or copy options.
    *   Progress bar.

2.  **Logic:**
    *   Laplacian variance calculation.
    *   Convert to grayscale first.
    *   Handle various image formats.

3.  **Deliverables:**
    *   `blur_detector.py`
    *   `run.bat`, `run.sh`
    *   `requirements.txt`

🧠 Prompt Decoding

• Laplacian Variance: A sharp image has high variation in the Laplacian (edges are clear). A blurry image has low variance (edges are mushy).

🛠️ Instructions

1. Install: pip install opencv-python
2. Copy Prompt → Run.
3. Adjust threshold based on your quality standards.