I followed the instruction above to successfully make the Quad-Camera Kit of four 16MP IMX519 cameras work with lowe speed driver on Raspberry Pi 4 model B in âmacroâ focus mode.
commands used:
libcamera-still -t 0 --autofocus-mode auto
libcamera-still -t 0 --autofocus-range macro
then, I tried to use âSwitch Composition Modeâ
The image was written normally only with the warning âERROR IPARPI cam_helper.cpp:249 Embedded data buffer parsing failedâ.
However, the captured image resolution ( 2328x1748-YUV420) was half of the max resolution of the 16MP IMX519 camera (4656 Ă 3496) .
Questions1:
Is there any command to get the max resolution of 16MP IMX519 with Arducam CamArray HAT Kit in Switch Composition Mode using âlibcamera-stillâ command?
Is there any other command to allow continuous switch channels and capture images at max resolution for each camera? My goal was to capture 4 images from each camera in max resolution,
Now I am using a loop to switch channels and " i2cset -y 10 0x24 0x24 0x02
libcamera-still -t 2000 -n -o test.jpg" to implement the function.
If there are better ways to achieve this goal, it will be highly appreciated.
The resolution is actually composed of four cameras. If you use dual-channel or single-channel, the corresponding resolution becomes 1/2 and 1/4 of the original
So with the Arducam CamArray HAT Kit, we can only achieve 1/4 of the 4656 (H) Ă 3496 (V), which is
1164Ă874 in single-channel using Switch Composition Mode?
Is the active resolution of 16MP IMX519, in fact, a 4M pixel camera?
I guess the limitation was still in the bandwidth of the Pi port.
Another follow-up question, does the driver can only use âimx519_kernel_driver_low_speedâ instead of âimx519_kernel_driverâ when using Arducam CamArray HAT Kit?
The title of the switch channel, please donât worry about the topic I share, both cameras use the same register on the switch channel.
I also want to clarify that a single channel will only allow you to save the full-resolution image of one camera. If you save four images, you need to switch to the corresponding channel for saving. Maybe the description on the product page is not very clear, I will let them revise the description.
I already tried with the commands like " i2cset -y 10 0x24 0x24 0x02
libcamera-still -t 2000 -n -o test.jpg" as mentioned in my former post,
the resolution of the images saved was not 16M. I will try again to follow your suggestions.
Follow-up questions:
1: What is the command to save the image at a specific resolution using âlibcameraâ?
Should we return the second set and get a new set? Or it is normal to have CamHelper for embedded data usage" error? We just want to make sure it will run stable and robust for our system.
This problem will not affect your use, the new version of libcamera will not parse the relevant data, there is a way to turn off the error
set environment variables:
export LIBCAMERA_NOTPARSE_EMBEDDED_DATA=1
Use âwidth --height to set the image resolution. Use âviewfinder-width --viewfinder-height to set preview resolution
Hi Edward, Following your suggestions, we already got the system working well.
System configuration (Arducam 16MP Autofocus Quad-Camera Kit and a Raspberry Pi 4 Computer Model B 8GB Single Board Computer).
We designed a program to let the Quad-Camera Kit to capture images at different locations by rotating the cameras together. Sometimes the images are out of focus and blurry, some are very sharp.
Is there a way to control the parameter of the autofocus of the Quad-Camera Kit?
What is the best way to capture images within focus and high quality? Factors like the distance to the object and illumination or delay for the camera to capture images.
Attached are two images captured by the system for comparison.
Your Pi will capture an image at a distance of 0.2m after 8 seconds.
Manual - The lens will never move spontaneously, but the "LensPosition" control can be used to move the lens
"manually". The units for this control are dioptres (1 / distance in metres), so that zero can be used to denote "inâ
finity". The "LensPosition" can be monitored in the image metadata too, and will indicate when the lens has
reached the requested location.