Multi_Adapter_Board V2.2 faulty images, missing scripts and examples

Hey! I recently bought the multi camera adapter board v2.2.

I had a few issues setting everything up, but now everything is working theoretically. In reality I am facing issues tho:

  1. If I run the “previewOpencv”, I get images of the board, but they are quite faulty. There are purple horizontal lines running through the images and parts of the images are completely black. How can I fix that (I reconnected all the cables, but it did not change a thing)

  2. How can I specify, that I am only using 3 cameras?

  3. For my project I idealy need to get the camera frames of each camera and provide them for another program. Is there any working simple example on how to get a video stream of the cameras like the “previewOpencv” to use in my program.

  4. Just a note: The included manual speaks about 2 scripts: “4cam_cv2.py” and “snapshot.py”. Both are not included in the github library.

Thanks in advance :slight_smile:

Just one example:

I think editing old messages is limited. (maybe you can change that to enable edit until another person replies)

 

I wanted to try if changing the camNum to 3 in the "AdapterBoard.py " works. If I try to run the “previewOpencv”-Python Version, there is an error printing “Write Error”, then the process does not respond.

Hello,

Thanks for your detail description. I will try my best to help you.

  1. If I run the “previewOpencv”, I get images of the board, but they are quite faulty. There are purple horizontal lines running through the images and parts of the images are completely black. How can I fix that (I reconnected all the cables, but it did not change a thing)

A: Do you open the VNC server? I also encountered the same problems before and found out that it is due the opening of VNC server.

 

  1. How can I specify, that I am only using 3 cameras?

A: For the C++ version, it will detected the sensor automatically. If the port exist sensor, the code will read the camera’s sensor ID to specify if the camera is exist.

  1. For my project I idealy need to get the camera frames of each camera and provide them for another program. Is there any working simple example on how to get a video stream of the cameras like the “previewOpencv” to use in my program.

A: We just have previewOpencv demo for users. You can refer to the previewOpencv about how to read image in your program.

  1. Just a note: The included manual speaks about 2 scripts: “4cam_cv2.py” and “snapshot.py”. Both are not included in the github library.

A: This is our old version code and Now we have updated our demo and delete the older code demo.

 

What’s more, using the python demo, if you set the camNum to 3, you should connect the camera to portA B and C.

Let me know if you need more help.

 

Adding to 1.) I was using a VNC server, I stopped it. Now the issue is gone.

 

i need to connect to the raspberry remotely. I just tried using a XRDP server. Now the issue is almost gone. Only cam C is having the same trouble with the corrupted image.

 

How can I connect to my pi remotely without these issues? I mean it must be an issue from Arducams side. Is there any plan on how fixxing it or is it a hardware issue?

  1. I am noticing a shift in colors (tint) between frames. How can I fix that/ensure that each frame has (almost) the same color.

Hello,

Thanks for contacting me. About the purple horizontal lines, This problem occurs when the remote network speed is not good. If the local display is normal, it means that the hardware is no problem.

About shift in colors between frames, I think it has something to do with our current hardware design. Due to the PI just has one CSI2 interface, so at the same time, only one sensor can be connected to the Pi, In another words, we can’t connect the four sensor to PI’s camera connector at the same time. So, we use a i2c multiplexer to choose each sensor. In this way,

The Pi’s ISP can’t adjust one sensor continuous. When the ISP get one sensor’s image and analyze exposure and awb, IT will set the result to another sensor, which will cause the four camera’s image has a little different.