Numpy array (gray scale) to QPixmap

robopassio

My input image is at grayscale, i.e., cv_image_array, which is a numpy array.
I can successfully save it at as

cv2.imwrite('gray.jpg', cv_image_array)

or display it as

cv2.imshow("Image", cv_image_array)
cv2.waitKey(1)

However, I want to display it directly on the UI

I did:

h, w = 320, 240
q_img = QImage(cv_image_array, w, h, w, QImage.Format_Grayscale8)
q_pixmap = QPixmap.fromImage(gray_img).scaled(h,w)
_widget.image1.setPixmap(q_pixmap)

However, the displayed image is broken and different from the one using cv2 commands
I tried a lot of time but I cannot find a solution.
Could someone please suggest some ideas to solve this?

SGaist

Hi and welcome to devnet,

Are you sure the image is stored as a single channel and not RGB with only gray colors ?

Chong Yoe Yat

Hi, i recently get into same problem also, what i did to solve is i changed the grayscale to color and replace frame.flatten() with frame.data and add bytesPerLine

I write it down here just in case someone need this. : - )

if len(cv_img.shape)<3:
    frame = cv2.cvtColor(cv_img, cv2.COLOR_GRAY2RGB)
else:
    frame = cv2.cvtColor(cv_img, cv2.COLOR_BGR2RGB)
h, w = cv_img.shape[:2]
bytesPerLine = 3 * w
qimage = QImage(frame.data, w, h, bytesPerLine, QImage.Format.Format_RGB888) 
#qimage = QImage(frame.flatten(), w, h, QImage.Format_RGB888)