Skip to content

Latest commit

 

History

History
executable file
·
88 lines (52 loc) · 3.69 KB

usage.md

File metadata and controls

executable file
·
88 lines (52 loc) · 3.69 KB
Raw

Folders and schema for working with the camera array

These directories either exist in the gamma repo or will be created during operation.

jot/ - unversioned directory, good place to save recordings retrieved from the array

notebooks/ - ipython notebooks

ccline/ - camera control and logging

deploy/ - software for setting up new nodes

mechanical/ - 3d model source and STLs

Nodes have the directories:

data/

  • r_D/ - data for recording ID r_D from this node.
  • nocollection/ - data recorded with no active recording ID.

Client commands

Get a sample image from each node

The sample command captures a single frame from a single node. To get one image now from node gamma1, use:

./scripts/run.py --client --gin_configs prod.gin --cmd sample --target_node_id gamma1

This command runs the CliRunner.camera_live_sample_cmd on each camera node in sequence. When complete it moves all the resulting images to jot/s_<datestamp>

j="jot/s_$(date +%s)/" ; mkdir $j ; for s in {1..6} ; do ./scripts/run.py --client --gin_configs prod.gin --cmd sample --target_node_id gamma${s} ; done ; mv sample-*.jpg $j

Collect imagery

Starting a collection will create a directory on each camera node at data/r_<datestamp> and save frames there.

./scripts/run.py --client --gin_configs prod.gin --cmd start

Issue the stop command to stop taking pictures.

./scripts/run.py --client --gin_configs prod.gin --cmd stop

On each node the Gamma camera server just runs the CliRunner.camera_video_cmd command from the supplied configuration file on that node. The provided configuration uses libcamera-vid.

Hardware UI

A subset of the functions are available from a display with buttons attached to one of the camera array nodes. A collection can be started or stopped and some stats can be viewed while collecting imagery. The UI delegates to the same client library, similar to the client commands above so that it's easy to turn any operation performed on the commandline into a menu action.

Running ui.py at boot

The example systemd service uiline.service is provided. Install it on the node where the UI hardware is present. Once the service is enabled it will run ui.py when that Raspberry Pi boots up. When making changes to ui.py itself the service has to be stopped or restarted.

Using the UI

When the ui is running with the OLED bonnet attached it shows a single menu item and a line of details relevant to that item. In general, pressing a button will activate the menu item which corresponds directly to running a Python function in ui.py. Since the code needs to be modified to suit the specifics of the camera array, it's best to read the code comments to get an idea of what the menu items actually do.

Retrieving data

Use rsync to fetch data from one or all nodes. The jot directory is the intended location for data to land. Use a pattern like r_collection-name, for example r_2022-09-01-hiking-trail.

r=r_collection-name; mkdir jot/${r} ; for h in {1..6} ; do echo ${h} ; rsync pi@gamma${h}:data/${r}/*.jpg jot/${r}/gamma${h}/ ; done

Retrieve imagery

The data from each node is stored on that node. Currently Gamma camera doesn't handle retrieving data from the array. The simplest way to to pull imagery from all nodes is probably rsync:

First find the name of the directory (it will be the same on all nodes):

ssh pi@gamma1 ls data/

The latest directory starting with r_, in this case it's r_1672602570. Run a command like this to copy files sequentially from each node to the workstation:

r=r_1672602570 ; mkdir jot/${r} ; for h in {1..6} ; do echo ${h} ; rsync [email protected].${h}:data/${r}/*.jpg jot/${r}/gamma${h}/ ; done