搜索一个二维数组内的子阵列（图像识别） [英] Searching a sub-array inside a 2D array (image recognition)

问题描述

从本质上讲，我有一个numpy的图像阵列，我试图找到它是否包含特定的RGB像素值的2×2块。因此，举例来说，如果我（简体）图像阵列是这样的：

Essentially, I have a numpy image array and I'm trying to find if it contains a 2x2 block of particular RGB pixel values. So, for example, if my (simplified) image array was something like:

A B C D E F

G H I J K L

M N O P Q R

S T U V W X

我想检查是否含有，说：

I am trying to check if it contains, say:

J K

P Q

我是pretty新numpy的，所以我倒是AP preciate有这方面的帮助，谢谢。

I'm pretty new to numpy so I'd appreciate any help on this, thanks.

推荐答案

如何此解决方案：

1）确定的大阵列中的小阵列的右上左侧元件的所有位置。

1) Identify all the locations of the upper-right left-hand element of the small array in the big array.

2）检查对应于每个给定元件的大阵列的片是完全一样的小阵

2) Check if the slice of the big array that corresponds to a every given element is exactly the same as the small array.

说，如果切片的上左角元件是5，我们会发现的5个地点的大阵列中，然后转到判断是否大阵列的切片左下角5是相同的小数组。

Say if the upper left-hand corner element of the slice is 5, we would find locations of 5 in the big array, and then go check if a slice of the big array to the bottom-left of 5 is the same as small array.

import numpy as np

a = np.array([[0,1,5,6,7],
              [0,4,5,6,8],
              [2,3,5,7,9]])

b = np.array([[5,6],
              [5,7]])

b2 = np.array([[6,7],
               [6,8],
               [7,9]])

def check(a, b, upper_left):
    ul_row = upper_left[0]
    ul_col = upper_left[1]
    b_rows, b_cols = b.shape
    a_slice = a[ul_row : ul_row + b_rows, :][:, ul_col : ul_col + b_cols]
    if a_slice.shape != b.shape:
        return False
    return (a_slice == b).all()

def find_slice(big_array, small_array):
    upper_left = np.argwhere(big_array == small_array[0,0])
    for ul in upper_left:
        if check(big_array, small_array, ul):
            return True
    else:
        return False

结果：

>>> find_slice(a, b)
True
>>> find_slice(a, b2)
True
>>> find_slice(a, np.array([[5,6], [5,8]]))
False

这篇关于搜索一个二维数组内的子阵列（图像识别）的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！