The ifm3d_ros2 package allows us to configure the O3R camera platform via two separate ways:
- Use ROS native service calls
- Use dump and config service proxies
Per camera head connected to the visual processing unit (VPU) of the O3R platform these services are available:
| Name | Service Definition | Description |
|---|---|---|
| Dump | ifm3d_ros2/camera/Dump | Dumps the state of the camera system as a JSON (formatted as a string) |
| Config | ifm3d_ros2/camera/Config | Provides a means to configure the VPU and Heads (imager settings), declaratively from a JSON (string) encoding of the desired settings. |
As you can see the two services Dump and Config are also part of this list.
Calling the native ROS service /ifm3d_ros2/camera/Dump for a certain camera will return a string containing with the camera configuration as a JSON string. Please notice the use of backslashes (\ before each ") to escape each upper quotation mark. This is necessary to allow us to keep the JSON syntax native to the underlying API (ifm3d).
Call this service via, e.g. for camera:
$ ros2 service call /ifm3d/camera/Dump ifm3d_ros2/srv/Dump
1637355550.468025 [0] ros2: using network interface enp0s31f6 (udp/192.168.0.10) selected arbitrarily from: enp0s31f6, wlp0s20f3, docker0, enx98fc84eebfc8
requester: making request: ifm3d_ros2.srv.Dump_Request()
response:
ifm3d_ros2.srv.Dump_Response(status=0, config='{"device":{"clock":{"currentTime":1581110268783361824},"diagnostic":{"temperatures":[],"upTime":19624000000000},"info":{"device":"0301","deviceTreeBinaryBlob":"tegra186-quill-p3310-1000-c03-00-base.dtb","features":{},"name":"","partNumber":"M03975","productionState":"AA","serialNumber":"000201234159","vendor":"0001"},"network":{"authorized_keys":"","ipAddressConfig":0,"macEth0":"00:04:4B:EA:9F:35","macEth1":"00:02:01:23:41:59","networkSpeed":1000,"staticIPv4Address":"192.168.0.69","staticIPv4Gateway":"192.168.0.201","staticIPv4SubNetMask":"255.255.255.0","useDHCP":false},"state":{"errorMessage":"","errorNumber":""},"swVersion":{"kernel":"4.9.140-l4t-r32.4+gc35f5eb9d1d9","l4t":"r32.4.3","os":"0.13.13-221","schema":"v0.1.0","swu":"0.15.12"}},"ports":{"port0":{"acquisition":{"framerate":10.0,"version":{"major":0,"minor":0,"patch":0}},"data":{"algoDebugConfig":{},"availablePCICOutput":[],"pcicTCPPort":50010},"info":{"device":"2301","deviceTreeBinaryBlobOverlay":"001-ov9782.dtbo","features":{"fov":{"horizontal":127,"vertical":80},"resolution":{"height":800,"width":1280},"type":"2D"},"name":"","partNumber":"M03969","productionState":"AA","sensor":"OV9782","sensorID":"OV9782_127x80_noIllu_Csample","serialNumber":"000000000395","vendor":"0001"},"mode":"experimental_autoexposure2D","processing":{"extrinsicHeadToUser":{"rotX":0.0,"rotY":0.0,"rotZ":0.0,"transX":0.0,"transY":0.0,"transZ":0.0},"version":{"major":0,"minor":0,"patch":0}},"state":"RUN"}}')
Below you can see an example on how to configure your camera via a ROS service call. The JSON string can be a partial JSON string. It only needs to follow basic JSON syntax. Please wrap the JSON string in a YAML syntax and use the field "json".
$ ros2 service call /ifm3d/camera/Config ifm3d_ros2/srv/Config "{json: '{\"ports\":{\"port0\":{\"mode\":\"standard_range4m\"}}}'}"
1637355711.266033 [0] ros2: using network interface enp0s31f6 (udp/192.168.0.10) selected arbitrarily from: enp0s31f6, wlp0s20f3, docker0, enx98fc84eebfc8
requester: making request: ifm3d_ros2.srv.Config_Request(json='{"ports":{"port2":{"mode":"standard_range4m"}}}')
response:
ifm3d_ros2.srv.Config_Response(status=0, msg='OK')
ifm3d-ros2 provides access to each camera parameter via the Dump and Config services exposed by the camera_node.
Additionally, command-line scripts called dump and config are provided as wrapper interfaces to the native API ifm3d. This gives a feel similar to using the underlying C++ API's command-line tool, from the ROS-independent driver except proxying the calls through the ROS network.
For example, to dump the state of the camera: (exemplary output from an O3R perception platform with one camera head connected is shown)
$ ros2 run ifm3d_ros2 dump
{
"device": {
"clock": {
"currentTime": 1581193835185485800
},
"diagnostic": {
"temperatures": [],
"upTime": 103190000000000
},
"info": {
"device": "0301",
"deviceTreeBinaryBlob": "tegra186-quill-p3310-1000-c03-00-base.dtb",
"features": {},
"name": "",
" partNumber": "M03975",
"productionState": "AA",
"serialNumber": "000201234160",
"vendor": "0001"
},
"network": {
"authorized_keys": "",
"ipAddressConfig": 0,
"macEth0": "00:04:4B:EA:9F:0E",
"macEth1": "00:02:01:23:41:60",
"networkSpeed": 1000,
"staticIPv4Address": "192.168.0.69",
"staticIPv4Gateway": "192.168.0.201",
"staticIPv4SubNetMask": "255.255.255.0",
"useDHCP": false
},
"state": {
"errorMessage": "",
"errorNumber": ""
},
"swVersion": {
"kernel": "4.9.140-l4t-r32.4+gc35f5eb9d1d9",
"l4t": "r32.4.3",
"os": "0.13.13-221",
"schema": "v0.1.0",
"swu": "0.15.12"
}
},
"ports": {
"port0": {
"acquisition": {
"framerate": 10,
"version": {
"major": 0,
"minor": 0,
"patch": 0
}
},
"data": {
"algoDebugConfig": {},
"availablePCICOutput": [],
"pcicTCPPort": 50010
},
"info": {
"device": "2301",
"deviceTreeBinaryBlobOverlay": "001-ov9782.dtbo",
"features": {
"fov": {
"horizontal": 127,
"vertical": 80
},
" resolution": {
"height": 800,
"width": 1280
},
"type": "2D"
},
"name": "",
"partNumber": "M03969",
"productionState": "AA",
"sensor": "OV9782",
"sensorID": "OV9782_127x80_noIllu_Csample",
"serialNumber": "000000000382",
"vendor": "0001"
},
"mode": "experimental_autoexposure2D",
"processing": {
"extrinsicHeadToUser": {
"rotX": 0,
"rotY": 0,
"rotZ": 0,
" transX": 0,
"transY": 0,
"transZ": 0
},
"version": {
"major": 0,
"minor": 0,
" patch": 0
}
},
"state": "RUN"
},
"port2": {
"acquisition": {
"exposureLong": 5000,
" exposureShort": 400,
"framerate": 10,
"offset": 0,
"version": {
"major": 0,
" minor": 0,
"patch": 0
}
},
"data": {
"algoDebugConfig": {},
"availablePCICOutput": [],
"pcicTCPPort": 50012
},
"info": {
"device": "3101",
"deviceTreeBinaryBlobOverlay": "001-irs2381c.dtbo",
"features": {
"fov": {
"horizontal": 105,
"vertical": 78
},
" resolution": {
"height": 172,
"width": 224
},
"type": "3D"
},
"name": "",
"partNumber": "M03969",
"productionState": "AA",
"sensor": "IRS2381C",
"sensorID": "IRS2381C_105x78_4x2W_110x90_C7",
"serialNumber": "000000000382",
"vendor": "0001"
},
"mode": "standard_range4m",
"processing": {
"diParam": {
"anfFilterSizeDiv2": 2,
"enableDynamicSymmetry": true,
"enableStraylight": true,
"enableTemporalFilter": true,
"excessiveCorrectionThreshAmp": 0.3,
"excessiveCorrectionThreshDist": 0.08,
"maxDistNoise": 0.02,
"maxSymmetry": 0.4,
"medianSizeDiv2": 0,
"minAmplitude": 20,
"minReflectivity": 0,
"mixedPixelFilterMode": 1,
"mixedPixelThresholdRad": 0.15
},
"extrinsicHeadToUser": {
"rotX": 1,
"rotY": 0,
"rotZ": 0,
"transX": 100,
"transY": 0,
"transZ": 0
},
"version": {
" major": 0,
"minor": 0,
"patch": 0
}
},
"state": "RUN"
}
}
}
Chaining together Linux pipelines works just as it does in ifm3d. For example, using a combination of dump and config to change the frame rate
from 5Hz to 10Hz of the single application on this particular camera would look like:
$ ros2 run ifm3d_ros2 dump | jq '.ports.port0.mode="standard_range2m"' | ros2 run ifm3d_ros2 config
$ ros2 run ifm3d_ros2 dump | jq .ports.port0.mode
"standard_range2m"
NOTE: If you do not have jq on your system, it can be installed with: $ sudo apt install jq
For the config command, one difference between our ROS implementation and the ifm3d implementation is that we only accept input on stdin. So, if you had a file with JSON you wish to configure your camera with, you would simply use the file I/O redirection facilities of your shell (or something like cat) to feed the data to stdin. For example, in bash:
$ ros2 run ifm3d_ros2 config < camera.json