2) olixVision™ Camera

(software patch > 1215)

Versions​ #

Product Name	Type	Generation	Product Number
Camera	USB-C Single IMU 3MP	01	CAM-U01X3M-S
Camera	USB-C Single IMU 3MP TPU	01	CAM-U01X3M-S-TP
Camera	USB-C Single IMU 3MP Global Shutter	01	CAM-U01X3M-S-G

Technical Specifications​ #

Feature Category	Feature Subcategory	Specification
Form Factor	Dimensions (W x H x D)	40mm x 40mm x 30mm
	Weight	89 grams
Processor Unit	Application Processor	Dual Cortex-A7 up to 800 MHz
	Real-Time Processor	Cortex-M4 MPU up to 200 MHz
	TPU AI Accelerator	4 Trillion Operations Per Second
Memory	On Chip (SoC)	512 MB RAM
	On Chip EEPROM (SoC)	512 Bytes x 8
	On SOM	64 GByte SD Flash
Sensors	Image Sensor	1/2.7″ OmniVision OV2710
	Max Resolution	1920(H)x1080(V) pixels
	Lens	2.1mm (L210) / 1.8mm (L180)
	Frame Rate	640×480 VGA @120fps, 1280×720 HD @60fps, 1920×1080 FHD @30fps
	IMU Sensor	6-Axis Automotive-Proven IMU
	IMU Range & Sensitivity	Accelerometer: 0.06 mg/LSB, Gyroscope: 0.004 dps/LSB
Connectivity	High-Speed Connectivity	1x Virtual Ethernet USB Type C @ 60 MBps
	Other I/O	1x User Switch, 3x User LEDs
Software	Yocto BSP	Available for batch purchases
	Linux Kernel	Linux 5.10
	Operating System	Debian 11
	Communication middleware	Apache Cyclone DDS
	Robotic Operating Systems	ROS Noetic Ninjemys, ROS 2 Humble Hawksbill
Power and Thermal	Power Consumption	USB Type C PD (15 W max)
	Voltage	PD 5.0v
	Max Current	3000mA
	Temperature Range	Commercial: 0°C to 85°C, Industrial: -40°C to +85°C

TPU​ #

A TPU, or Tensor Processing Unit, is a type of AI accelerator specifically designed by Google to accelerate machine learning tasks. TPUs are application-specific integrated circuits (ASICs) that have been optimized for the efficient execution of tensor operations, which are fundamental to deep learning and other machine learning algorithms.

The primary goal of a TPU is to enhance the performance and energy efficiency of machine learning workloads, allowing AI models to be trained and executed more quickly and with lower power consumption than traditional CPU or GPU-based hardware. This is achieved through a combination of specialized hardware components and optimizations tailored to the unique requirements of machine learning tasks.

Enable / Disable TPU acceleration​ #

In order to enable the tpu chip do

./opt/olive/script/enable_tpu.sh

and to disable the tpu do

./opt/olive/script/disable_tpu.sh

to verify the TPU is enabled you can do

lsusb

The result will be:

Bus 001 Device 004: ID 1a6e:089a Global Unichip Corp.
Bus 001 Device 003: ID 32e4:9230 HD USB Camera HD USB Camera
Bus 001 Device 002: ID 1a40:0101 Terminus Technology Inc. Hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub

Which the “Global Unichip Corp” is the TPU chip.

TPU Apps​ #

All the python3 based TPU examples are located in /home/olive/apps. To run an object detection example using TPU after enabling the tpu in the previus step run

python3 /home/olive/apps/packages/src/app_tpu_python/tpu_python/app_node.py

Note that you have to change the topic name of the app subscriber to your current device topic name in oder to subscribe to the image. In order to to this you have to edit the app_node.py

After running this example the camera will detect objects based on Coral object detection example.

General pipeline diagram

For more information about Google® coral please visit:

https://coral.ai/examples

For other python example source projects please visit:

https://github.com/olive-robotics/olv_camera_tpu_playground_py

General Setup​ #

Follow the Quick Start for Olive Robotics robot modules to connect the device and start using it.

Quick Setup​ #

Quick start guide for the olive® Camera module:

1. Check if the data is alreaddy published to your system using:ros2 topic listyou should be able to see all the topics published from the module. For exampe for IMU module you will see:
   • /olive/imu/id001/image/camera_info
   • /olive/imu/id001/image/compressed
   • /olive/imu/id001/imu
   • /olive/imu/id001/magnetometer
   • /olive/imu/id001/status
   • /parameter_events
   • /rosout

The number id001 is your device’s default namespace.

2. Visualize the data using:
   1. Embedded Web Interface
   2. rqt (https://docs.ros.org/en/humble/Tutorials/Beginner-CLI-Tools/Introducing-Turtlesim/Introducing-Turtlesim.html#install-rqt)
      • Visualize the topics:
      • Plot the data from the IMU messages:
      • Visualize the Image:
      • Change the parameters of the DCM:
   3. rviz2 (https://github.com/ros2/rviz).
      • Visualize the IMU data in Rviz2
3. Visualize the data and change the parameters like IP, Topic Name, etc using the embedded web interface.

ROS Topics and Services​ #

Topic Name	Message Type	Type	Description
…/image/camera_info	sensor_msgs/CameraInfo	Publisher	Camera image information
…/image/compressed	sensor_msgs/CompressedImage	Publisher	Camera image
…/imu	sensor_msgs/Imu	Publisher	Messured acc/gyro/quaternion
…/magneticfield	sensor_msgs/MagneticField	Publisher	Messured Magnetic field.
…/status	diagnostic_msgs/DiagnosticStatus	Publisher	Device status.

Advanced Settings​ #

The device allows certain parameters to be changed at runtime. To get an overview of all changeable parameters use ros2 param list. To change a parameter use ros2 param set /dcm_camera <parameter> <new_value>

Camera​ #

Parameter	Type	Range Min	Range Max	Default	Description
frequency	int	0	120	30	Camera publish rate
brightness	double	0.0	1.0	0.5	Camera brightness parameter
contrast	double	0.0	1.0	0.5	Camera contrast parameter
saturation	double	0.0	1.0	0.5	Camera saturation parameter
gamma	double	0.0	1.0	0.5	Camera gamma parameter
whitebalance	double	0.0	1.0	0.5	Camera whitebalance parameter
resolution	string	“320×240”	“1920×1080”	“640×480”	Camera image resolution

IMU​ #

Parameter	Type	Range Min	Range Max	Default	Description
filter_frequency	int	0	1200	100	IMU filter frequency
filter_gain	int	0	1	0.2	IMU filter gain
frequency_imu	int	0	1200	100	IMU publish rate
frequency_mag	int	0	100	100	Magnetometer publish rate

System​ #

Parameter	Type	Range Min	Range Max	Default	Description
frequency	int	0	10	10	System status publish rate

Advanced setup​ #

Camera Calibration​ #

• Install Camera Calibration Parser, Camera Info Manager and Launch Testing Ament Cmake

sudo apt install ros-humble-camera-calibration-parsers
sudo apt install ros-humble-camera-info-manager
sudo apt install ros-humble-launch-testing-ament-cmake

• The image_pipeline need to be built from source in your workspace with:

git clone – b humble git@github.com:ros-perception/image_pipeline.git

• A large checkerboard with known dimensions. This tutorial uses a 7×9 checkerboard with 20mm squares. Calibration uses the interior vertex points of the checkerboard, so an “8×10” board uses the interior vertex parameter “7×9” as in the example below. The checkerboard with set dimensions can be downloaded from here.

https://calib.io/pages/camera-calibration-pattern-generator

• A well-lit area clear of obstructions and other check board patterns
• A monocular camera publishing images over ROS

ros2 run camera_calibration cameracalibrator --size 7x9 --square 0.02 --ros-args -r image:=/my_camera/image_raw -p camera:=/my_camera

Explanation of the required parameters:

Camera Name:

-c, --camera_name
        name of the camera to appear in the calibration file

Chessboard Options:

You must specify one or more chessboards as pairs of --size and--square options.

  -p PATTERN, --pattern=PATTERN
                    calibration pattern to detect - 'chessboard','circles', 'acircles','charuco'
  -s SIZE, --size=SIZE
                    chessboard size as NxM, counting interior corners (e.g. a standard chessboard is 7x7)
  -q SQUARE, --square=SQUARE
                    chessboard square size in meters

ROS Communication Options:

 --approximate=APPROXIMATE
                    allow specified slop (in seconds) when pairing images from unsynchronized stereo cameras
 --no-service-check
                    disable check for set_camera_info services at startup

Calibration Optimizer Options:

 --fix-principal-point
                    fix the principal point at the image center
 --fix-aspect-ratio
                    enforce focal lengths (fx, fy) are equal
 --zero-tangent-dist
                    set tangential distortion coefficients (p1, p2) to
                    zero
 -k NUM_COEFFS, --k-coefficients=NUM_COEFFS
                    number of radial distortion coefficients to use (up to
                    6, default 2)
 --disable_calib_cb_fast_check
                    uses the CALIB_CB_FAST_CHECK flag for findChessboardCorners

This will open a calibration window which highlight the checkerboard.

As the checkerboard is moved around the 4 bars on the calibration sidebar increases in length. When all then the 4 bars are green and enough data is available for calibration the CALIBRATE button will light up. Click it to see the results. It takes around the minute for calibration to take place.

After the calibration is completed the save and commit buttons light up. And you can also see the result in terminal.

Press the save button to see the result. Data is saved to “/tmp/calibrationdata.tar.gz”

default factory calibrated parameters for 320×240 are as follows:

image_width: 320
image_height: 240
camera_name: narrow_stereo
camera_matrix:
  rows: 3
  cols: 3
  data: [353.88557,   0.     , 160.14916,
        0.     , 353.7126 ,    117.27734,
        0.     ,   0.     ,   1.     ]
distortion_model: plumb_bob
distortion_coefficients:
  rows: 1
  cols: 5
  data: [-0.428635, 0.167437, 0.001243, 0.004107, 0.000000]
rectification_matrix:
  rows: 3
  cols: 3
  data: [1., 0., 0.,
         0., 1., 0.,
         0., 0., 1.]
projection_matrix:
  rows: 3
  cols: 4
  data: [319.03568,   0.     , 161.28226,   0.    ,
        0.     , 334.55365, 116.9051 ,   0.     ,
        0.     ,   0.     ,   1.     ,   0.     ]

place the ost.yaml in the current folder of the camera:

/opt/olive/config/current/camera

Now restart the camera and it should use your new calibration parameters for camera_info topic.

​Downloads​ #

Type	Format	Version	Link
3D Models
	OBJ	1.0	Download
	DAE	1.0	Download

Supported Camera Lenses​ #

Lens	Focal Length (mm)	Max Aperture	Sensor Size	Features	Predicted FOV (°)
2.8-12mm 1:1.4 IR	2.8-12	f/1.4	1/2.7″	Manual focus, zoom	~28° – ~90°
Far-view	2.6	f/2.6	1/4″	N/A	~55°
Low-distortion Lens	3.6	N/A	1/2.7″	N/A	~60°
Fisheye Lens	10 (estimated)	f/2.8 (estimated)	1/2.3″ (estimated)	N/A	187°

Mechanical Details​ #