做直方圖均衡case2時,出現 ValueError: all the input arrays must have same number of dimensions錯誤
做直方圖均衡case2時,出現 ValueError: all the input arrays must have same number of dimensions,不太明白哪邊出錯了,謝謝!
```
import cv2
import numpy as np
img_path = '/lena.png'
img = cv2.imread(img_path, cv2.IMREAD_COLOR)
# case 1
# 每個 channel 個別做直方圖均衡
equalHist_by_channel_B = cv2.equalizeHist(img[...,0])
equalHist_by_channel_G = cv2.equalizeHist(img[...,1])
equalHist_by_channel_R = cv2.equalizeHist(img[...,2])
# 組合經過直方圖均衡的每個 channel
img_bgr_equal = np.hstack((equalHist_by_channel_B,equalHist_by_channel_G,equalHist_by_channel_R))
cv2.imshow("img_bgr_equal",img_bgr_equal)
#=======================================================================
# case 2 - 轉換 color space 後只對其中一個 channel 做直方圖均衡
img_hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
img_hsv_equal = cv2.equalizeHist(img_hsv[...,0])
# 組合圖片 + 顯示圖片
cv2.imshow("img_bgr_equal",img_bgr_equal)
cv2.imshow("img",img)
cv2.imshow("img_hsv_equal",img_hsv_equal)
img_bgr_equalHist = np.hstack((img, img_bgr_equal, img_hsv_equal))
while True:
# 比較 (原圖, BGR color space 對每個 channel 做直方圖均衡, HSV color space 對明度做直方圖均衡)
cv2.imshow('bgr equal histogram', img_bgr_equalHist)
k = cv2.waitKey(0)
if k == 27:
cv2.destroyAllWindows()
break
```
```
ValueError
Traceback (most recent call last) in
25 cv2.imshow("img_hsv_equal",img_hsv_equal)
26--->
27img_bgr_equalHist = np.hstack((img, img_bgr_equal, img_hsv_equal))
28whileTrue:
29# 比較 (原圖, BGR color space 對每個 channel 做直方圖均衡, HSV color space 對明度做直方圖均衡)
/anaconda3/lib/python3.7/site-packages/numpy/core/shape_base.py in hstack(tup)
338 return _nx.concatenate(arrs,0)
339else:-->
340 return _nx.concatenate(arrs,1)
341
342 ValueError: all the input arrays must have same number of dimensions
```
回答列表
-
2019/11/28 上午 09:07K.F贊同數：1不贊同數：0留言數：0
按照訊息
np.hstack((img, img_bgr_equal, img_hsv_equal))
你這三張圖size是不一樣的
建議你印出size比較看看
-
2019/12/07 下午 03:26Chen-Ming Yang贊同數：0不贊同數：0留言數：0
Hello,
根據 K.F. 的敘述補充一下
當我們想要把矩陣疊起來的時候, 必須要確保他們的 size 都一樣
也就是 img.shape == img_bgr_equal.shape == img_hsv_equal.shape