Why does gstreamer capture image incorrectly?

Question

Hello,

I have a Jetson Nano device with a Raspberry Pi v2.1-camera and use gstreamer and OpenCV to capture images with it. Still, I thought this question would be better preserved here because I mainly think it is a software issue. The point is: I use following Python-script to capture my image:

import cv2
import numpy as np
import matplotlib.pyplot as plt
from scipy.misc import imshow

def gstreamer_pipeline (capture_width=3280, capture_height=2464, display_width=1280, display_height=720, framerate=21, flip_method=0) :   
    return ('nvarguscamerasrc ! ' 
    'video/x-raw(memory:NVMM), '
    'width=(int)%d, height=(int)%d, '
    'format=(string)NV12, framerate=(fraction)%d/1 ! '
    'nvvidconv flip-method=%d ! '
    'video/x-raw, width=(int)%d, height=(int)%d, format=(string)BGRx ! '
    'videoconvert ! '
    'video/x-raw, format=(string)BGR ! appsink'  % (capture_width,capture_height,framerate,flip_method,display_width,display_height))


if __name__ == '__main__':
    cap = cv2.VideoCapture(gstreamer_pipeline(flip_method=0), cv2.CAP_GSTREAMER)
    if cap.isOpened():
        ret_val, img = cap.read()
        img = np.flipud(img)
        img = np.fliplr(img)
        imshow(img)
        cv2.imwrite("test.jpg", img)
    else:
        print("Unable to open camera.")

My problem is I am not sure if it captures color values correctly. When I use imshow(img) from SciPy to display the image I just took, it looks like this: (I do not have enough reputation to insert it, so please see this link)

That is strange because it is very blue-toned. After I use cv2.imwrite("test.jpg", img) to save it, it looks normal again: (Here the another image)

I do not know if I defined something wrong in the gstreamer-pipeline or what else could cause the error. It is important for my program to capture colors correctly and since I do not want to save an image and then load it just to get the colors right (bad in terms of speed and CPU-power), I would be very glad if someone could help me out.

Thanks for the answers in advance!

Answer 1

As mentioned by HansHirse your color space is wrong.

You could alter your gstreamer pipeline to output RGB

format=(string)RGB    

Or convert from BGR to RGB within oopencv

im_rgb = cv2.cvtColor(im_cv, cv2.COLOR_BGR2RGB)