图像处理之矩阵二维卷积

Diko

于 2018-08-07 19:12:59 发布

阅读量634

点赞数

分类专栏：图像处理文章标签：图像处理二维卷积矩阵卷积矩阵

本文链接：https://blog.csdn.net/Kedongyu_/article/details/81478265

版权

本文详细探讨了图像处理中的二维卷积概念，通过矩阵运算阐述了卷积在图像处理中的作用，包括滤波、特征提取等关键步骤，帮助读者深入理解这一核心技术。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

# -*- coding: utf-8 -*-
"""
Created on Wed May 02 15:08:01 2018

@author: Diko
"""

import numpy 
from skimage import io ,data

def conv2d_v0010(f_, g_):
    g = g_[::-1, ::-1]
    f = f_
    h = numpy.vstack([numpy.zeros(f_.shape[1] + g_.shape[1] - 1)] * (g_.shape[0] + f_.shape[0] -1))
    
    f_up_and_dowm = numpy.vstack([numpy.zeros(f.shape[1])]*g.shape[0])
    f = numpy.vstack((f_up_and_dowm,f,f_up_and_dowm))
    
    f_left_and_right = numpy.vstack([numpy.zeros(g.shape[1])] * f.shape[0])
    f = numpy.hstack((f_left_and_right, f, f_left_and_right))
    
    for i in xrange(f_.shape[0] + g_.shape[0] - 1):
        for j in xrange(f_.shape[1] + g_.shape[1] -1):
            h[i, j]=(f[i+1:i+1+g_.shape[0],j+1:j+1+g_.shape[1]] *g).sum()
            
    print "f=\n", f
    print "g=\n", g
    print "h=\n",