提取图像像素数据存入数组时遇ValueError错误求助

问题场景 #

需要从图像列表中提取64x64x3尺寸的像素数据，将所有值存入数组X，要求每张图像以单独一列的形式表示，原代码运行时触发ValueError。

原代码 #

Y = np.zeros((1,m))
X = np.zeros((ROWS*COLS*CHANNELS, m))

for i,image_file in enumerate(train_images):
    if 'cat' in image_file.lower():
        Y[0,i] = 1
    else:
        Y[0,i] = 0
    
    img = cv2.imread(image_file, cv2.IMREAD_COLOR)
    img_resized = cv2.resize(img, (ROWS, COLS),interpolation=cv2.INTER_CUBIC)
    X[0, i] = np.squeeze(img_resized.reshape((ROWS*COLS*CHANNELS,1)))

报错信息 #

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
TypeError: only size-1 arrays can be converted to Python scalars

The above exception was the direct cause of the following exception:

ValueError                                Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_17808\256590185.py in <module>
     10     img = cv2.imread(image_file, cv2.IMREAD_COLOR)
     11     img_resized = cv2.resize(img, (ROWS, COLS),interpolation=cv2.INTER_CUBIC)
---> 12     X[0, i] = np.squeeze(img_resized.reshape((ROWS*COLS*CHANNELS,1)))

ValueError: setting an array element with a sequence.

错误原因 #

代码中X[0, i]试图对数组X的第0行第i列这个标量位置赋值，但img_resized.reshape(...)展开后是长度为64*64*3=12288的一维数组，二者维度不匹配，导致赋值失败。

修正方案 #

将赋值语句的索引从X[0, i]改为X[:, i]，表示对X的第i列整体赋值，匹配展开后的图像像素数组维度：

Y = np.zeros((1,m))
X = np.zeros((ROWS*COLS*CHANNELS, m))

for i,image_file in enumerate(train_images):
    if 'cat' in image_file.lower():
        Y[0,i] = 1
    else:
        Y[0,i] = 0
    
    img = cv2.imread(image_file, cv2.IMREAD_COLOR)
    img_resized = cv2.resize(img, (ROWS, COLS),interpolation=cv2.INTER_CUBIC)
    # 修正索引，用[:,i]选中整列，同时用flatten简化展开操作
    X[:, i] = img_resized.flatten()

额外优化建议 #

• 用img_resized.flatten()替代reshape(...) + np.squeeze()，代码更简洁高效；
• 提前确认ROWS=64、COLS=64、CHANNELS=3，避免因变量定义错误导致维度不匹配；
• OpenCV默认以BGR格式读取图像，若需要RGB格式，可添加img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)转换。

内容的提问来源于stack exchange，提问作者Justin Wilker