I WANT TO USE OPENCV IN THE PYTHON MR./MS. PROGRAM OF CS/EG SERIES CAMERAS (IN THE CASE OF BAYER)

The Python program for CS/EG series cameras is not coded to rely on OpenCV to be compatible with OpenCV. It is a generic code that does not depend on any specific library.

 

In the following, we will use “GrabImage.py” as an example.

 

Python-OpenCV treats image data as numpy arrays, so you need to adapt to its data format. In GrabImage.py, we use MV_CC_GetOneFrameTimeout() in work_thread() to get the image, and the image data is passed in pData, which is equivalent to a pointer in C.

 

As it is, it cannot be handled by OpenCV, so we need to convert this pData to a numpy array. Follow the steps below to do the conversion.

 

1) Add the following two lines of import statements:

• import numpy as np
• import cv2

2) __main__ ret = cam. Before the MV_CC_StartGrabbing () line, add the following line to set the image data format output from the camera.

• ret = cam.MV_CC_SetEnumValueByString(“PixelFormat”,”BayerRG8“)

The BayerRG8 part varies depending on the camera, so open MVS, connect the camera, and select the appropriate string from Feature Tree → Image Fromat Control → Pixel Format.

 

Display in MVS	Corresponding string
Bayer RG 8	BayerRG8
Bayer GR 8	BayerGR8
Bayer BG 8	BayerBG8
Bayer GB 8	BayerGB8

 

 

 

３）Inside work_thread(), add code to convert pData to numpy arrays. (I’ll also add two lines, cv2.imshow() and cv2.waitKey(), so that you can check the image.) 

def work_thread(cam=0, pData=0, nDataSize=0):     stFrameInfo = MV_FRAME_OUT_INFO_EX()     memset(byref(stFrameInfo), 0, sizeof(stFrameInfo))     while True :         ret = cam.MV_CC_GetOneFrameTimeout(pData, nDataSize, stFrameInfo, 1000)         if ret == 0:             print (“get one frame: Width[%d], Height[%d], nFrameNum[%d]”  % (stFrameInfo.nWidth, stFrameInfo.nHeight, stFrameInfo.nFrameNum))             image = np.asarray(pData._obj)             image = image.reshape((stFrameInfo.nHeight, stFrameInfo.nWidth))             image = cv2.cvtColor(image, cv2.COLOR_BayerBG2BGR)             cv2.imshow(“show”, image)             cv2.waitKey(1)         else:             print (“no data[0x%x]” % ret)         if g_bExit == True:             break

 

cv2. Enter a character string corresponding to the character string you just entered in the COLOR_BayerBG2BGR part.

 

	Corresponding string
BayerRG8	cv2.COLOR_BayerBG2BGR
BayerGR8	cv2.COLOR_BayerGB2BGR
BayerBG8	cv2.COLOR_BayerRG2BGR
BayerGB8	cv2.COLOR_BayerGR2BGR

 

The notation of the second argument of cv2.cvtColor() does not match the intuitive Bayer sequence.

This is an OpenCV specification.

 

Specifications can be found below.