How to Effectively Determine HSV Color Boundaries for Object Detection using cv::inRange?

Choosing Color Boundaries for Object Detection with cv::inRange (OpenCV)

When utilizing the cv::inRange function for color detection, selecting appropriate upper and lower HSV boundaries is crucial. This article addresses the question of how to effectively determine these boundaries based on a specific color of interest.

Background

HSV (Hue, Saturation, Value) is a color space commonly used in image processing. The HSV model represents colors as three components:

• Hue (H): Represents the color shade (e.g., red, blue).
• Saturation (S): Measures the amount of color present in the shade (0-1).
• Value (V): Represents the brightness of the color (0-255).

Choosing Boundaries

Determining proper HSV boundaries is based on the specific color being detected. Here's a step-by-step guide:

1. Determine Hue:

   • Use a color picker tool to identify the HSV values of the object of interest.
   • Note that different scales might be used for HSV values depending on the application.

2. Adjust Hue Range:

   • Account for slight variations in hue by adjusting the range around the identified value.
   • For example, if the hue is 22 (out of 179), a range of (11-33) could be appropriate.

3. Set Saturation and Value Ranges:

   • Use a reasonable range for saturation (e.g., 50-255).
   • For value, choose a range that includes the expected brightness of the object.

4. Consider Format:

   • Ensure that the HSV conversion is appropriate for your image format.
   • For example, OpenCV uses BGR, not RGB for image representation.

Example

Let's consider the example of detecting an orange lid in an image.

1. HSV Values:

   • Using a color picker, we obtain an HSV value of (22, 59, 100).

2. Adjusted Boundaries:

   • Hue range: (11-33)
   • Saturation range: (50-255)
   • Value range: (50-255)

3. Python Code:

   import cv2
   import numpy as np
   
   ORANGE_MIN = np.array([11, 50, 50], np.uint8)
   ORANGE_MAX = np.array([33, 255, 255], np.uint8)
   
   # Read and convert image
   img = cv2.imread('image.png')
   hsv_img = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
   
   # Detect orange using inRange
   mask = cv2.inRange(hsv_img, ORANGE_MIN, ORANGE_MAX)
   # Display mask
   cv2.imshow('Mask', mask)
   cv2.waitKey(0)

The above is the detailed content of How to Effectively Determine HSV Color Boundaries for Object Detection using cv::inRange?. For more information, please follow other related articles on the PHP Chinese website!