Using CSI camera
Introduction
The Linaro Debian and/or OpenEmbedded releases for the DragonBoard 410c include drivers for:
- OV5645 camera sensor
- Qualcomm Camera Subsystem (CAMSS)
- Qualcomm Camera Control Interface (CCI)
OV5645
The OV5645 is a 5MP MIPI CSI2 camera sensor. The driver supports three sensor modes:
- 2592x1944 15fps (full frame)
- 1920x1080 30fps (cropped from full frame)
- 1280x960 30fps
The driver was accepted upstream and was merged in the Linux Kernel v4.12. It is also backported in the Linaro releases kernel branches.
CAMSS
The Camera Subsystem hardware found on 8x16 processors consists of:
- 2 CSIPHY modules. They handle the Physical layer of the CSI2 receivers. A separate camera sensor can be connected to each of the CSIPHY module;
- 2 CSID (CSI Decoder) modules. They handle the Protocol and Application layer of the CSI2 receivers. A CSID can decode data stream from any of the CSIPHY. Each CSID also contains a TG (Test Generator) block which can generate artificial input data for test purposes;
- ISPIF (ISP Interface) module. Handles the routing of the data streams from the CSIDs to the inputs of the VFE;
- VFE (Video Front End) module. Contains a pipeline of image processing hardware blocks. The VFE has different input interfaces. The PIX input interface feeds the input data to the image processing pipeline. The image processing pipeline contains also a scale and crop module at the end. Three RDI input interfaces bypass the image processing pipeline. The VFE also contains the AXI bus interface which writes the output data to memory.
The driver implements the V4L2 subdev interface and Media Controller API. With the goal to model the hardware links between the modules and to expose a clean, logical and usable interface, the driver is split into V4L2 sub-devices as follows:
- 2 CSIPHY sub-devices - each CSIPHY is represented by a single sub-device;
- 2 CSID sub-devices - each CSID is represented by a single sub-device;
- 2 ISPIF sub-devices - ISPIF is represented by a number of sub-devices equal to the number of CSID sub-devices;
- 4 VFE sub-devices - VFE is represented by a number of sub-devices equal to the number of the input interfaces (3 RDI and 1 PIX).
The driver supports:
- Input from a camera sensor via CSIPHY
- Generation of test input data by the TG in CSID
- RDI interface of VFE - raw dump of the input data to memory. Supported formats:
- YUYV/UYVY/YVYU/VYUY (packed YUV 4:2:2)
- MIPI RAW8 (8bit Bayer RAW)
- MIPI RAW10 (10bit packed Bayer RAW)
- MIPI RAW12 (12bit packed Bayer RAW)
- PIX interface of VFE
- Format conversion of the input data
- Supported input formats:
- YUYV/UYVY/YVYU/VYUY (packed YUV 4:2:2).
- Supported output formats:
- NV12/NV21 (two plane YUV 4:2:0);
- NV16/NV61 (two plane YUV 4:2:2).
- Supported input formats:
- Scaling. Configuration of the VFE Encoder Scale module for downscalling with ratio up to 16x.
- Cropping. Configuration of the VFE Encoder Crop module.
- Format conversion of the input data
- Concurrent and independent usage of two data inputs - could be camera sensors and/or TG.
The driver was accepted upstream and is expected in the Linux Kernel v4.14. It is also backported in the Linaro releases kernel branches.
CCI
The CCI is a I2C controller dedicated for camera control.
The driver is implemented as an I2C adapter driver and will be upstreemed.
Enable camera
The Linaro release is configured with no camera and users with camera are expected to configure the DTS file accordingly. Please check commit “dts: Disable camera sensors in dtsi” in the kernel. This patch assumes that two OV5645 camera sensors are connected on the board (one per CSI2 channel) and disables them. To enable a camera please revert (part of) this patch.
Basic usage
Make sure that the following package is installed:
sudo apt-get install v4l-utils
To ensure that the sensor is properly connected and detected, you can inspect the output of the following command:
sudo media-ctl -d /dev/media0 -p
If everything is ok, you should see something like this:
- entity 87: ov5645 4-003b (1 pad, 1 link)
type V4L2 subdev subtype Sensor flags 0
device node name /dev/v4l-subdev10
pad0: Source
[fmt:UYVY8_2X8/1920x1080 field:none
crop:(0,0)/1920x1080]
-> "msm_csiphy0":0 [ENABLED,IMMUTABLE]
Direct dump to memory
You need to configure the Media Controller pipeline: link CSIPHY to CSID, CSID to ISPIF, ISPIF to VFE. Then configure formats on all entities in the pipeline. For direct dump to memory (RDI channels) this looks like this:
sudo media-ctl -d /dev/media0 -l '"msm_csiphy0":1->"msm_csid0":0[1],"msm_csid0":1->"msm_ispif0":0[1],"msm_ispif0":1->"msm_vfe0_rdi0":0[1]'
sudo media-ctl -d /dev/media0 -V '"ov5645 4-003b":0[fmt:UYVY8_2X8/1920x1080 field:none],"msm_csiphy0":0[fmt:UYVY8_2X8/1920x1080 field:none],"msm_csid0":0[fmt:UYVY8_2X8/1920x1080 field:none],"msm_ispif0":0[fmt:UYVY8_2X8/1920x1080 field:none],"msm_vfe0_rdi0":0[fmt:UYVY8_2X8/1920x1080 field:none]'
At this point the pipeline should be configured and ready to be used by any application that supports V4L2. For example, you can use GStreamer to take a JPEG picture:
gst-launch-1.0 v4l2src device=/dev/video0 num-buffers=1 ! 'video/x-raw,format=UYVY,width=1920,height=1080' ! jpegenc ! filesink location=image01.jpg
Or you can use GStreamer to show a live preview from the camera:
gst-launch-1.0 v4l2src ! 'video/x-raw,format=UYVY,width=1920,height=1080' ! glimagesink
If you have a second camera sensor and intend to use it concurrently then link and configure another pipeline which includes the second camera and the unused entities. Use a v4l2 application the same way only pointing the correct video device node used in this pipeline.
Format conversion
Pipeline configuration for the format conversion looks like this:
sudo media-ctl -d /dev/media0 -l '"msm_csiphy0":1->"msm_csid0":0[1],"msm_csid0":1->"msm_ispif0":0[1],"msm_ispif0":1->"msm_vfe0_pix":0[1]'
Format configuration for NV16/NV61 output:
sudo media-ctl -d /dev/media0 -V '"ov5645 4-003b":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csiphy0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csid0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_ispif0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":1[fmt:UYVY8_2X8/1280x960 field:none]'
Format configuration for NV12/NV21 output (the format on msm_vfe0_pix source pad - for NV12/NV21 must be UYVY8_1_5_X8 - must be set explicitly):
sudo media-ctl -d /dev/media0 -V '"ov5645 4-003b":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csiphy0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csid0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_ispif0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":1[fmt:UYVY8_1_5X8/1280x960 field:none]'
And similar Gstreamer pipeline for a JPEG picture:
gst-launch-1.0 v4l2src device=/dev/video3 num-buffers=1 ! 'video/x-raw,format=NV12,width=1280,height=960,framerate=30/1' ! jpegenc ! filesink location=image02.jpg
Scale
Format configuration for NV12/NV21 with downscale ratio 2x. The compose element on msm_vfe0_pix sink pad defines the output size from the scaler. Syntax is (left,top)/widthxheight and only width and height are valid as this is scaling only. Downscaling with up to 16x ratio is supported:
media-ctl -d /dev/media0 -V '"ov5645 4-003b":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csiphy0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csid0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_ispif0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":0[fmt:UYVY8_2X8/1280x960 field:none compose:(0,0)/640x480],"msm_vfe0_pix":1[fmt:UYVY8_1_5X8/640x480 field:none]'
Crop
Format configuration for NV12/NV21 with crop of the bottom right corner. The crop element on msm_vfe0_pix source pad defines the cropped area. Syntax is (left,top)/widthxheight and all fields are valid:
media-ctl -d /dev/media0 -V '"ov5645 4-003b":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csiphy0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csid0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_ispif0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":1[fmt:UYVY8_1_5X8/320x240 field:none crop:(960,720)/320x240]'
Scale + crop
Format configuration for NV12/NV21 with downscale ratio 2x and crop of the center area. Scaling is done first and then cropping (scaler module is in front of the crop module in the hardware pipeline of the VFE):
media-ctl -d /dev/media0 -V '"ov5645 4-003b":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csiphy0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_csid0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_ispif0":0[fmt:UYVY8_2X8/1280x960 field:none],"msm_vfe0_pix":0[fmt:UYVY8_2X8/1280x960 field:none compose:(0,0)/640x480],"msm_vfe0_pix":1[fmt:UYVY8_1_5X8/320x240 field:none crop:(160,120)/320x240]'
Video record pipeline
Starting with 17.06 release a video recording GStreamer pipeline is supported involving the camera and video encoder. Currently this has the following limitations:
- A GStreamer video encoder plugin is needed which is not a part of the release. To enable this the user must patch and rebuild
gstreamer1.0-plugins-good
package - this is recommended for advanced users only. Instructions are available in the release notes.
Example GStreamer pipeline for video recording:
gst-launch-1.0 -e v4l2src device=/dev/video3 ! video/x-raw,format=NV12,width=1280,height=960,framerate=30/1 ! v4l2h264enc extra-controls="controls,h264_profile=4,video_bitrate=2000000;" ! h264parse ! mp4mux ! filesink location=video.mp4
Using CSID Test Generator
If you do not have any camera sensor, it is possible to use the internal CSID Test Generator (TG). To enable it download and compile the yavta
tool from here:
git clone git://git.ideasonboard.org/yavta.git
cd yavta
make
Then, enable the test generator:
./yavta --no-query -w '0x009f0903 1' /dev/v4l-subdev2
Then, configure the pipeline:
sudo media-ctl -d /dev/media0 -l '"msm_csid0":1->"msm_ispif0":0[1],"msm_ispif0":1->"msm_vfe0_rdi0":0[1]'
sudo media-ctl -d /dev/media0 -V '"msm_csid0":1[fmt:UYVY8_2X8/1920x1080 field:none],"msm_ispif0":0[fmt:UYVY8_2X8/1920x1080 field:none],"msm_vfe0_rdi0":0[fmt:UYVY8_2X8/1920x1080 field:none]'
Finally, you can use any v4l2 application, such as GStreamer.