clock-time-extraction-opencv

Chapter 1: Solving Methods

1.1 Introduction

Topic: Retrieving the time from a photo of a clock
Requirements:

• Input: At least 5 photos containing analog wall clocks with varying content (different clock types, 1-2 hour time differences, non-coincident clock hands).
• Output:
  • Draw rectangular frames around hour, minute, and second hands in each photo, with distinct colors. Automated frame drawing, not manual.
  • Output time as hours:minutes:seconds for each photo.

1.2 Methods

1.2.1 Overall

The method uses cv2.HoughCircles or cv2.findContours to detect clocks, then cv2.HoughLinesP to detect hands. Mathematical methods calculate hand rotation angles, leading to time inference.

1.2.2 Detailed Description

1. Step 1: Image Preprocessing

   • Resize image for complexity reduction.
   • Convert BGR to HSV color space for brightness and color handling.
   • Invert HSV color to enhance contrast.
   • Apply Otsu thresholding to HSV channel V for binary image.
   • Gaussian blur to reduce noise.

2. Step 2: Clock Detection

   • Use cv2.HoughCircles or cv2.findContours.
   • Identify largest circle or contour as clock.

3. Step 3: Detect Line Segments

   • Apply Canny filter to find edges.
   • Use HoughLinesP to detect straight lines.

4. Step 4: Group Lines

   • Find close, nearly parallel lines.
   • Group lines within clock radius and similar angles.

5. Step 5: Detect Clock Hands

   • Find endpoints of line segments.
   • Create clock hands from farthest point to clock center.
   • Determine hand thickness.

6. Step 6: Determine Hand Types

   • Identify hour, minute, and second hands.

7. Step 7: Draw Frames

   • Use cv2.rectangle to frame clock hands.
   • Use cv2.putText to label hands.

8. Step 8: Determine Rotation Angle

   • Calculate angle between clock hands and center.

9. Step 9: Calculate Time

   • Calculate time from hand angles.

10. Step 10: Draw Time on Image

• Draw time on Image

---

TEST

To test the clock time extraction functions on 10 images, run the following command: python clock_time_extraction.py images/ adjusted_images/

This command will process the clock images located in the images/ directory and save the adjusted clock images in the adjusted_images/ directory. Make sure to replace images/ and adjusted_images/ with the actual paths to your input and output directories.

---

Author: TrietCS
Course: Introduction to Digital Image Processing