Introudction

This the MIPI camera software SDK for Raspberry pi platform, which allow you connect global shutter cameras and upto 18MP cameras on RPI board.

Now the supported MIPI camera modules are below:

0.3MP OV7251 Monochrome Global Shutter
1MP OV9281 Monochrome Global Shutter
2MP OV2311 Monochrome Global Shutter
13MP IMX135 Color Rolling Shutter
16MP IMX298 Color Rolling Shutter
18MP AR1820 Color Rolling Shutter

Video Demo

Prerequisites

Enable the camera

Install support package

sudo apt-get update && sudo apt-get install libzbar-dev libopencv-dev

sudo apt-get install python-opencv

Download and install the SDK library

$ pi@raspberrypi:~ $ git clone https://github.com/ArduCAM/MIPI_Camera.git

pi@raspberrypi:~ $ cd MIPI_Camera/RPI/

pi@raspberrypi:~/MIPI_Camera/RPI $ make install

pi@raspberrypi:~/MIPI_Camera/RPI $ chmod +x enable_i2c_vc.sh

pi@raspberrypi:~/MIPI_Camera/RPI $ ./enable_i2c_vc.sh

[Notice] For Pi4 platform

cd /tmp

wget https://project-downloads.drogon.net/wiringpi-latest.deb

sudo dpkg -i wiringpi-latest.deb

Then click Y/y to reboot now

Compile the Examples

pi@raspberrypi:~ $ cd MIPI_Camera/RPI/

pi@raspberrypi:~/MIPI_Camera/RPI $ make clean && make

How to use release demos? (use the imx298 as an example)

arducamstill

This is a comprehensive application, user can use it to preview, get image in different format and fine tuning the image quality.

get help

pi@raspberrypi:~/MIPI_Camera/RPI $ ./arducamstill -?

Previdew in mode 1 and enable awb enable ae

pi@raspberrypi:~/MIPI_Camera/RPI $ ./arducamstill -t 0 -m 1 -awb 1 -ae 1

- From the print message, you can see the current mode framerate .etc

 Current mode: 1, width: 1280, height: 960, pixelformat: pBAA, desc: (null)
[Framerate]: 02 pfs, [Exposure]: 2680, [Focus]: 0190,[Rgain]: 0140, [Bgain]: 0070

Preview 5 seconds, then get a image in jpg format

pi@raspberrypi:~/MIPI_Camera/RPI $ ./arducamstill -t 5000 -m 1 -e jpg -o test.jpg

Detail parameters

-t, --timeout   : Time (in ms) before takes picture and shuts down (if not specified, loop)
-q, --quality   : Set jpeg quality <0 to 100>
-m, --mode      : Set sensor mode
-awb, --autowhitebalance        : Enable or disable awb
-ae, --autoexposure     : Enable or disable ae
-rgain, --awbrgain      : Set R channel gian vaue <0 to 65535>
-bgain, --awbbgain      : Set B channel gian vaue <0 to 65535>
-o, --capture   : usd to get one frame
-e, --encoding  : Encoding to use for output file (jpg, bmp, gif, png，raw)
-?, --help      : This help information

Fine tuning the exposure focus and gain using the keyboard.
After getting a good result, you can set the curruent value using arducam_set_control API. For the detail APIs, please refer to

https://github.com/ArduCAM/MIPI_Camera/blob/master/RPI/arducam_mipicamera.h

ISP demo

Arducam release ISP demo depend on opencv

cd /MIPI_Camera/RPI/ISP

 python preview.py

opencvGui

For the convenience of users to see the effect of fine-tuning intuitively, arduacm release a GUI application

pi@raspberrypi:~/MIPI_Camera/RPI $ ./opencvGui

Toolbar introduction

-Quit					: close the application
-AWB					: enable and disable auto white balance function
-AE						: enable and disable auto exposure function
-Reset					: reset all parameters 
-Snapshot				: get one frame image 
-exposure step 1 		: fine-tuning exposure time 
-focus step 1 			: fine-tuning focus value
-awb rgain compensation : fine-tuning awb rgain compensation
-awb bgain compensation : fine-tuning awb bgain compensation

Press Ctrl + C to exit the application

list_format

This demo is used to check all resolution the sensor you are using supports

pi@raspberrypi:~/MIPI_Camera/RPI $ ./list_format

capture This demo is used to capture one frame image with the special encoder format we set .

pi@raspberrypi:~/MIPI_Camera/RPI $ ./capture

If I choose mode 0 The default encoder is jpeg, if you want to change others, just opne the capture.c file and change here:

Notice: if you want to get the image with BMP or PNG encoder, you shuold set the encoder time to more. I advise you set 12000 in the capture.c file.

How to use Arducam stereo hat? (use the ov9281 as an example)

Hardware setup
Run the ./list_format to check which mode is used for stereo hat.

pi@raspberrypi:~/MIPI_Camera/RPI $ ./list_format

From the print message, we can see the mode 7-11 ared used for stereo hat. I choose the mode7 as a demo

pi@raspberrypi:~/MIPI_Camera/RPI $ ./preview_setMode 7

Optional Settings

Edit /boot/config.txt file. Find gpu_mem=xxx line Modify gpu_mem size with proper size, recommend to use

gpu_mem=160 for 13MP camera board,

gpu_mem=180 for 16MP or higher camera board.

Running the Examples

preview_setMode Example

$ ./preview_setMode [mode]

In the preview_setMode example, it will demo how to do preview the Specified mode listed from ./list_format command.

Capture Example

$ ./capture [mode]

In the capture.c example, it will capture diffferent resolution in with different encoder, support JPEG BMP PNG.

$ ./capture_raw

In the capture_raw.c example, it will capture diffferent resolution none interpolation raw format images, especially useful for monochrome sensors.

$ ./raw_callback

In the raw_callback.c example, it is callback version of capture_raw example.

Video Recording Example

$ ./video

In the video.c example, it will record the video in H246 format.

Camera Control Query Example

$ ./list_format

In the list_format.c example, it will list camera supported resolution and control functions.

Sensor Register Read/Write Example

$ ./read_write_sensor_reg

In the read_write_sensor_reg.c example, it illustrates how to directly read/write sensor registers. This example might need to be modifed according to the correct sensor register address.

OpenCV Example

$ ./capture2opencv

In the capture2opencv.cpp example, it converts YUV data to OpenCV Mat format, and displays as is.

Gstreamer Example

In the video2stdout.c example, it outputs H.264 video stream to stdout, and uses gstreamer to push the stream to PC.

Example 1:

Raspberry pi side command:

$ ./video2stdout | nc -l -p 5000

PC side command: (x.x.x.x is your Raspberry Pi IP address)

$  gst-launch-1.0  -v tcpclientsrc  host=x.x.x.x  port=5000 ! decodebin ! autovideosink

Example 2:

Raspberry pi side command: (x.x.x.x is your Raspberry Pi IP address)

$ ./video2stdout | gst-launch-1.0 -v fdsrc ! h264parse !  rtph264pay config-interval=1 pt=96 ! gdppay ! tcpserversink host=x.x.x.x port=5000

PC side command: (x.x.x.x is your Raspberry Pi IP address)

$  gst-launch-1.0 -v tcpclientsrc host=x.x.x.x port=5000 ! gdpdepay ! rtph264depay ! avdec_h264 ! autovideosink sync=false

QR Code Detection Example

$ ./qrcode_detection <exposure_value>

In the qrcode_detection.cpp example, it illustrates how to use global shutter camera like OV7251 or OV9281 to detect QR code using OpenCV. To run this demo you have to install the dependence

sudo apt-get update && sudo apt-get install libzbar-dev libopencv-dev

Dual Camera Demo

$ ./preview-dualcam
or
$ ./capture-dualcam

In the preview-dualcam.c examle, it illustrates how to open the two camera ports on Raspberry pi compute module at the same time for preview. And the capture-dualcam.c examle, it illustrates how to do capture from each camera port on Raspberry pi compute module by switching between them.

A camera_interface struct should be constructed according to your hardware wiring.

For example camera port 0 is using sda_pin 28, scl_pin 29, led_pin 30, shutdown_pin 31, and camera port 1 is using sda_pin 0, scl_pin 1, led_pin 2, shutdown_pin 3.

More information about the compute module wiring please check : https://www.raspberrypi.org/documentation/hardware/computemodule/cmio-camera.md

OV9281 External Trigger Demo

In the ov9281_external_trigger.c example, you can see how to synchronize the camera to the external trigger events.

Python Wrapper and Examples

The arducam_mipicamera.py script is a wrapper for the libarducam_mipicamera.so dynamic library. To use this script you need to pre-install libarducam_mipicamera.so. All python examples are in the Python Folder.

capture_raw10_opencv.py demo

This demo used to get raw10 image. Then unpack it to raw16. Display it using openCV

Dependency

sudo pip install v4l2

sudo pip install numpy

sudo apt-get install python-opencv

Utility

How to playback the H264 file

Compile hello_video.bin

$ cd /opt/vc/src/hello_pi && ./rebuild.sh

Play H264 file

$ /opt/vc/src/hello_pi/hello_video/hello_video.bin test.h264

How to view RAW data

In the utils folder, there are two python script to read and display RAW image.

mipi_raw10_to_jpg.py is used to display color RAW image.

mono_to_jpg.py is used to display monochrome RAW iamge.

Trouble shooting

Error message open camera status = -1

Didn't enable the camera, see Prerequisites->Enable the camera.

Didn't enable i2c_vc, see Prerequisites->Enable i2c_vc.

Camera cable loose.

Run the script camera_i2c from utils folder and send output message to support@arducam.com

sudo chmod +x camera_i2c rpi3-gpiovirtbuf

./camera_i2c

e.g imx298

GPU memory is too small

You should set the gpu_mem to bigger. Open the config.txt file. Then reboot

ArduCAM/PX_IMC01_RPI