Question

我正在尝试从单个numpy数组中的文件夹中读取多个图像，以将它们提供给我的深度学习模型。当我打印形状时，我得到一个错误消息，即没有类型的对象没有形状。这意味着图像没有被opencv（对于jpg）/ tifffile（对于.tif）读取我的代码如下

import numpy as np
import cv2 as cv
import tifffile as tiff

filepath1 = "D:/Programs/transnet/dataset/train"
filepath2 = "D:/Programs/transnet/dataset/train_label"
img = np.ndarray((cv.imread(filepath1 + "*.jpg")))
clt = np.ndarray((tiff.imread(filepath2 + "*.tif")))

print(img.shape)
print(clt.shape)

我已经尝试过glob.glob，但是它不起作用。我期望一个4维的rgb图像数组

Answer 1

您可以使用os模块执行文件遍历操作，尤其是如下的os.listdir和os.path.join

import os

def get_all_images(folder, ext):

    all_files = []
    #Iterate through all files in folder
    for file in os.listdir(folder):
        #Get the file extension
        _,  file_ext = os.path.splitext(file)

        #If file is of given extension, get it's full path and append to list
        if ext in file_ext:
            full_file_path = os.path.join(folder, file)
            all_files.append(full_file_path)

    #Get list of all files
    return all_files

filepath1 = "D:/Programs/transnet/dataset/train"
filepath2 = "D:/Programs/transnet/dataset/train_label"

#List of all jps and tif files
jpg_files = get_all_images(filepath1, 'jpg')
tif_files = get_all_images(filepath2, 'tif')

现在，一旦有了所有文件的列表，就可以遍历列表并打开图像。

是否有python函数可将文件夹中的多个图像读取到单个数组中

1 个答案: