How to extract texture from images using Python

How to use Python to extract texture from images

Introduction: Texture extraction is an important technology in image processing, which can extract texture features from images for Image analysis, classification, matching and other tasks. As a popular programming language, Python has a wealth of image processing libraries and tools, such as PIL, opencv-python, etc. This article will introduce how to use Python to extract texture from images, and attach code examples.

1. Install and import necessary libraries
Before starting, we need to install and import some necessary libraries, including PIL, opencv-python and numpy. They can be installed using pip:

$ pip install Pillow
$ pip install opencv-python
$ pip install numpy

The code to import the required libraries is as follows:

import cv2
import numpy as np
from PIL import Image

2. Reading and displaying images
First, we need to read an image to be processed , and displayed. Here, we selected a natural scenery picture and saved it as "texture.jpg".

# 读取图片
image = Image.open("texture.jpg")
# 显示图片
image.show()

3. Convert the image to a grayscale image
Texture extraction is generally performed on a grayscale image, so we need to convert the read color image into a grayscale image.

# 转换为灰度图像
gray_image = image.convert("L")
# 显示灰度图像
gray_image.show()

4. Calculate the gray level co-occurrence matrix (GLCM)
The gray level co-occurrence matrix is ​​a common method for texture analysis, which can describe the gray level changes and spatial relationships between pixels.

In Python, we can use the opencv-python library (cv2) to calculate the gray level co-occurrence matrix. The specific steps are as follows:

1. First, we need to convert the grayscale image into a numpy array for subsequent processing.

# 将灰度图像转换为numpy数组
gray_array = np.array(gray_image)

2. Next, we can use the cv2.calcHist() function in the cv2 library to calculate the gray level co-occurrence matrix. This function requires input of a grayscale image and some parameters, such as the number of grayscale levels, pixel intervals, image height and width, etc.

# 计算灰度共生矩阵
glcm = cv2.calcHist([gray_array], [0], None, [256], [0, 256])

3. Finally, we can normalize the gray-level co-occurrence matrix to facilitate subsequent analysis and feature extraction.

# 规范化灰度共生矩阵
glcm /= glcm.sum()

5. Texture feature extraction
After obtaining the gray level co-occurrence matrix, we can use it to extract some common texture features, such as energy, contrast, correlation Correlation and homogeneity, etc.

The following are several common texture feature calculation methods:

1. Energy (energy): used to measure the clarity of texture. The calculation formula is that the energy is equal to the sum of the squares of each element in the gray-level co-occurrence matrix.

# 计算能量（energy）
energy = np.sum(glcm ** 2)
print("能量：", energy)

2. Contrast: Used to measure the contrast between gray levels in a texture. The calculation formula is that the contrast is equal to the sum of each element in the gray-level co-occurrence matrix multiplied by the pixel difference at the corresponding position.

# 计算对比度（contrast）
contrast = np.sum(glcm * np.abs(np.arange(256) - np.arange(256)[:,None]))
print("对比度：", contrast)

3. Correlation: used to measure the linear correlation of textures. The calculation formula is that the correlation is equal to the sum of the products of each element in the gray-level co-occurrence matrix multiplied by the pixels at the corresponding positions.

# 计算相关性（correlation）
correlation = np.sum(glcm * (np.arange(256) - np.mean(glcm)) * (np.arange(256)[:,None] - np.mean(glcm))) / (np.std(glcm) ** 2)
print("相关性：", correlation)

4. Homogeneity: Used to measure the smoothness between gray levels in a texture. The calculation formula is that uniformity is equal to the sum of each element in the gray-level co-occurrence matrix divided by the pixel difference at the corresponding position.

# 计算均匀性（homogeneity）
homogeneity = np.sum(glcm / (1 + np.abs(np.arange(256) - np.arange(256)[:,None])))
print("均匀性：", homogeneity)

6. Summary
This article introduces how to use Python to extract texture from images and how to calculate common texture features. By using PIL and opencv-python, two commonly used image processing libraries, we can easily read and display images, convert images into grayscale images, and calculate the grayscale co-occurrence matrix of the image. After obtaining the gray-level co-occurrence matrix, we also introduced how to calculate common texture features, including energy, contrast, correlation, and uniformity.

Texture analysis is an important research direction in computer vision and plays an important role in image analysis, classification, retrieval and other tasks. I hope this article will be helpful to readers who learn and apply texture extraction technology.

The above is the detailed content of How to extract texture from images using Python. For more information, please follow other related articles on the PHP Chinese website!