Skip to content

Test setup

How to set up InSiSo Visual Detector on a ctrlX CORE for a typical usage and test scenario. It is written for Bosch Rexroth's app validation and works for any commissioning check. The two Postman collections it refers to, standard-scenario1.json and standard-scenario2.json, come with the app's validation package.

What you need

Item Needed for
ctrlX CORE X3 (arm64 snap), ctrlX CORE X5, X7 or ctrlX CORE virtual (amd64 snap), or a ctrlX IPC with the NVIDIA extension (+nvidia snap), on ctrlX OS 4.x Everything
The Runtime licence, installed or as a trial Everything
A PC with a browser and Postman Running the scenarios
A GenICam / GigE Vision or USB / UVC camera, the Camera Connector and Color Detection licences Scenario 2 only

Scenario 1 needs no camera and no model.

Set up the device

  1. Install the snap for the device: see Installation and Which snap.
  2. Install the licences, or start their trials, in Settings ▸ Apps & Licenses.
  3. Open InSiSo Visual Detector ▸ Diagnostics. The Licence card lists each function as unlocked or names its licence. The first chip, ctrlX Data Layer, shows ✔.
  4. In Postman, import both collections. Set the collection variables baseURL (for example https://192.168.1.1), username and password to the device's address and a ctrlX OS user.

Scenario 1: no peripherals

standard-scenario1.json. It changes no setting and can be run any number of times.

  1. Signs in with the ctrlX OS user (identity manager) and keeps the token.
  2. Reads ready, message and busy under insiso/visual-detector/.
  3. Runs the status commands through command and reads each answer in command-result: list-commands, scan-cameras, performance-status, storage-list.
  4. Writes postman-scenario-1 to trigger and reads what it published: picture/id, picture/trigger, picture/ok, picture/error, color/count, object/count, coordinates/count and json.
  5. Calls the REST API the pages use: licences, health, performance, cameras, models and the detection history.

Expected: every call answers HTTP 200. Each command's answer names its command. With no camera, ready is FALSE and message says why; the trigger publishes picture/ok FALSE with the reason in picture/error.

Scenario 2: one camera

standard-scenario2.json. Connect one GigE Vision or USB camera to the device first.

  1. Signs in.
  2. scan-cameras finds the camera and keeps its address in the variable cameraSource. To pick another camera, set cameraSource by hand.
  3. connect-camera connects it as camera 1. set-trigger-mode 1,software makes it take a picture only when triggered. camera-status 1 reads its state.
  4. create-color adds red to the camera's color detector, color-1.
  5. Writes trigger twice and reads the results of each picture: picture/id, picture/trigger, picture/ok, color/count, color/name, color/x-px, color/y-px, object/count and json.
  6. detector-status color-1 reads the detector's state.
  7. set-trigger-mode 1,free-run leaves the camera running and the detector evaluating every picture: the state a commissioned cell runs in.

Expected: ready is TRUE from step 4 on. Each trigger raises picture/id by one and echoes its value in picture/trigger. With a red part in view, color/name lists red and color/x-px and color/y-px give its centre.

Power cycle

Power-cycle the device without signing in. Within three minutes the app runs, the camera is reconnected with its trigger mode, the detectors run, and ready is TRUE. picture/id carries on past its last value. Nobody needs to press anything.

Clean up

Write disconnect-camera 1 to command, or press Disconnect on Camera Connector. Reset colors on Color Detection removes the red added by scenario 2.