OpenCV - numpy copy后图像未显示

Question

我正在尝试阅读＆＃34; .jpg＆＃34; python中的图像使用cv2.imread（）成为一个numpy数组。代码对我来说似乎很好但是当我使用plt.imshow（）显示图像时，我看到在将图像复制到数组后图像无法正确显示。不知道我在这里做错了什么。我错过了什么？

以下是代码：

import cv2
import os
import pandas as pd
import numpy as np

import matplotlib.pyplot as plt
%matplotlib inline

# Get current working directory
cwd = os.getcwd()
print(cwd)

# Read the csv file into a data frame
driving_log_df = pd.read_csv('driving_log.csv')
print(driving_log_df.shape)

X_data_set = np.empty([len(driving_log_df['Center'])*3, 32, 32, 3])
Y_data_set = np.empty(len(driving_log_df['Center'])*3)
print(X_data_set.shape)
print(Y_data_set.shape)

# Path to images 
images_path = cwd + "/IMG"
print(images_path)

# Index 
index = 0

for file in os.listdir(images_path):
    image = cv2.imread(os.path.join(images_path, file), cv2.IMREAD_COLOR)        
    image = cv2.resize(image, (32, 32))

    # OpenCV reads images in the BGR format, convert them into RGB
    image_rgb =  cv2.cvtColor(image, cv2.COLOR_BGR2RGB)

    # Copy the image 
    X_data_set[index] = np.copy(image_rgb)

    # Display images
    if(index==0):
        figure1 = plt.figure()
        plt.imshow(image_rgb)        
        figure2 = plt.figure()
        plt.imshow(X_data_set[0])

    index = index + 1

还附有正在显示的图像。

Answer 1

cv2中的轴顺序是bgr。所以你需要np.transpose到rgb。