Put Greyscale Images into one List - (Error: (-215) scn == 3 || scn == 4 in function ipp_cvtColor)

Question

I want to read every images in a folder into one list and then greyscale every image. All the greyscale images should be in one list as well. Currently, I am able to read images to one list, but I am not able to put all the greyscale images into one list.

Here is my code:

import cv2
import numpy as np
from os import listdir
from os.path import isfile, join

mypath = 'jpl_thomas' 
onlyfiles = [f for f in listdir(mypath) if isfile(join(mypath,f)) ] 

images = np.empty(len(onlyfiles), dtype=object) 
for n in range(0, len(onlyfiles)):  
    images[n] =cv2.imread(join(mypath,onlyfiles[n]))

frames = np.zeros(len(images)) 
for i in range(0, len(images)):
    frames[i] = cv2.cvtColor(images[i],cv2.COLOR_BGR2GRAY)

The error i get is as the following:

File "main.python", line 21, in <module>
    frames[i] = cv2.cvtColor(images[i],cv2.COLOR_BGR2GRAY) 
    cv2.error: /Users/jenkins/miniconda/1/x64/conda-bld/conda_1486588158526/work/opencv-3.1.0/modules/imgproc/src/color.cpp:7456: 
    error: (-215) scn == 3 || scn == 4 in function ipp_cvtColor

I tried to debug small parts. frames[0] = cv2.cvtColor(images[0],cv2.COLOR_BGR2GRAY) returns the right value. But if I use for loop, the error above occurs.

thewaywewere · Accepted Answer · 2017-04-30T00:10:47.837

The following code reads all the color image files in mypath and converts them to gray scale images at the same time. color_images[] and gray_images[] are the two list holding the color and gray images.

from os import listdir
from os.path import isfile, join
import numpy as np
import cv2

mypath = './imageFiles/' 
onlyfiles = [fname for fname in listdir(mypath) if isfile(join(mypath,fname))] 

color_images = np.empty(len(onlyfiles), dtype=object)
gray_images = np.empty(len(onlyfiles), dtype=object)

for n in range(0, len(onlyfiles)):  
    color_images[n] =cv2.imread(join(mypath,onlyfiles[n]))
    gray_images[n] =cv2.imread(join(mypath,onlyfiles[n]),0)

# plot images in list[] for checking
for i in range(len(color_images)):
    plt.figure(i)
    plt.subplot(121), plt.imshow(color_images[i][:,:,::-1]), plt.title('color') 
    plt.subplot(122), plt.imshow(gray_images[i], cmap='gray'), plt.title('grayscale')
plt.show()

The sample plots below are the image files put into _images[] which consist of the famous "Lena" and "Scenetext" image files under mypath.

Take note that, with OpenCV:

Color image is stored in numpy as (height[0:n], width[0:m], channels[B,G,R]) and
Gray scale image is only with one channel as (height[0:n], width[:0:m]).
cv2.imread(filename[,flags]) converts color image to gray scale if the flags is set to 0. For more flags, see "read an image".

Hope this help.

Your code works! Thank you for your detailed explanation. I have a further question. Why frames[i] = cv2.cvtColor(images[i],cv2.COLOR_BGR2GRAY) does not work in the 'for loop'? Why cvColor function works when I test one image separately, but does not work when I tried to put them in one list? What is the meaning of the error I got? Thank you! — Judie, Apr 30 '17 at 03:40
@Judie The error tells the input image `image[i]` should have 3`[B,G,R]` or 4`[B,G,R,A]` channels before applying the function `cv2.cvtColor()`. You may check the image shape to ensure the input type is correct before applying the function by `height, width, channels = image[i].shape`. — thewaywewere, Apr 30 '17 at 10:14
Thanks, I see. When I checked the image shape, I get ''NoneType' object has no attribute 'shape''. So I know that image[i] are nontype objects. I could not apply cvtColor function directly to image[i]. — Judie, Apr 30 '17 at 15:58

Put Greyscale Images into one List - (Error: (-215) scn == 3 || scn == 4 in function ipp_cvtColor)

1 Answers1